A pipeline monitoring system and method based on a magnetic induction coil

Abstract

The application discloses a pipeline monitoring system and method based on a magnetic induction coil. The pipeline monitoring system comprises at least two transceiving coils, at least two transceiving terminals and an upper computer. The transceiving terminals are electrically connected with the transceiving coils one by one, the transceiving terminals are used for outputting an excitation electric signal to the corresponding transceiving coil or receiving an induced electromagnetic signal from the corresponding transceiving coil, and the magnetic field at the transceiving coil is distributed along the circumference of the pipeline. The transceiving terminals are electrically connected with the upper computer, and the upper computer is used for analyzing the information carried in all electromagnetic signals by using a pre-trained classification model to obtain a pipeline monitoring result. The application directly induces the change of the peripheral environment of the pipeline caused by leakage by using the annular spiral coil, extracts features from the frequency spectrum of the received signal, can quickly detect the leakage condition and position, has a simple structure, low cost and application value in the energy industry, agriculture and other fields.

Description

Technical Field

[0001] This invention relates to the field of electronic technology, and more specifically to pipeline monitoring systems and methods. Background Technology

[0002] The waste of resources and industrial pollution caused by liquid pipeline leaks is an extremely important issue. Most liquid pipelines are located deep underground, and leaks are relatively inconspicuous, which poses a challenge to timely and accurate detection of leaks.

[0003] Traditional pressure sensors and flow meters are very expensive in large-diameter pipeline systems; traditional acoustic sensors are affected by external interference such as noisy environments and ground vibrations; and infrared imaging solutions cannot perform accurate quantitative analysis.

[0004] Magnetic induction technology, through non-propagating near-field electromagnetic coupling, avoids the attenuation problem of electromagnetic signals caused by underground soil, making it superior for monitoring long-distance transport pipelines. However, previous research has mainly focused on metal pipelines and used traditional loop coils. Plastic pipelines are almost unsuitable for using such coils; lossy liquids and polymer pipes cause power transmission between coils to attenuate.

[0005] Therefore, a pipeline monitoring system based on the principle of magnetic induction and a toroidal spiral coil is proposed to address the above problems.

[0006] Patent application CN113781749A discloses an improvement in the hardware structure of a pipeline leak location and early warning device in an underground utility tunnel. It includes: an optical fiber protective shell, a device outer shell, a platinum film layer, a fixed base, left and right resonant chambers, a first optical fiber, a second optical fiber, three fiber optic temperature sensors, two fiber optic strain sensors, and left and right sound collectors. The device uses two fiber optic temperature sensors to sense the ambient temperature as a reference temperature, one fiber optic temperature sensor to sense the temperature of the platinum film layer to detect pipeline leaks, and two fiber optic strain sensors to sense the sound of leaks from the left and right sides of the pipeline. The signals are connected to a photodetector via optical fibers, and the parameters are demodulated in real time by a high-speed fiber optic demodulator. This invention uses fiber optic sensors, resulting in a complex design; furthermore, a leak in one part of the pipeline can cause the entire monitoring system to malfunction, requiring the replacement of the entire sensor network.

[0007] Patent application CN218036691U discloses an online intelligent monitoring instrument for underground pipelines, including a pipeline; a settlement monitoring device is installed on the pipeline; the settlement monitoring device includes a first semi-circular slip ring, a second semi-circular slip ring, and a connecting rod; the first semi-circular slip ring is fitted around the outer ring of the pipeline cross-section; a settlement monitoring instrument is installed on the arc-shaped top of the first semi-circular slip ring; a first positive electrode plate is fixedly connected to the bottom of a fixing plate; a second positive electrode plate is fixedly connected to the top of a support plate; a negative electrode plate is installed at the end of the connecting rod near the first semi-circular slip ring; a leakage monitoring structure is installed on the second semi-circular slip ring; the leakage monitoring structure includes a special monitoring instrument; a special monitoring instrument is installed on the top of the second semi-circular slip ring; through the settlement monitoring device and the leakage monitoring structure, the condition of the underground pipeline can be accurately and timely monitored. This invention uses a special monitoring instrument for pipeline leakage monitoring, but does not specify the specific sensor type. However, general humidity sensors are often not sensitive enough and do not have the characteristic of real-time monitoring. Summary of the Invention

[0008] To address the shortcomings of existing technologies, this invention provides a pipeline monitoring system and method based on magnetic induction coils. This invention employs a toroidal spiral coil to directly sense changes in the pipeline's external environment caused by leaks, extracting features from the spectrum of the received signal to achieve rapid detection of pipeline leaks and their location.

[0009] The technical solution adopted in this invention is as follows:

[0010] (I) A pipeline monitoring system based on magnetic induction coils

[0011] The pipeline monitoring system includes at least two transceiver coils, at least two transceiver terminals, and a host computer. Each transceiver terminal is electrically connected to a corresponding transceiver coil. The transceiver terminal outputs an excitation electrical signal to its corresponding transceiver coil or receives an induced electromagnetic signal from its corresponding transceiver coil. Each transceiver coil receives an excitation electrical signal from its corresponding transceiver terminal and then generates a coupled electromagnetic field distribution outside the pipeline, or outputs an induced electromagnetic signal from its own location to its corresponding transceiver terminal. The magnetic field of the transceiver coil is distributed circumferentially along the pipeline. Each transceiver terminal wirelessly connects to the host computer to input the information carried in the received induced electromagnetic signal. The host computer processes the received information using a pre-trained classification model to obtain the pipeline monitoring results.

[0012] The pipeline monitoring system also includes relay coil groups, with a relay coil group set between every two adjacent transceiver coils. Each relay coil group includes several relay coils. All transceiver coils and relay coils are fitted onto the outside of the pipeline and arranged at equal intervals along the pipeline direction. The first and last coils arranged sequentially along the pipeline direction are both transceiver coils. All transceiver coils and relay coils are circular spiral coils, and the axis of the circular spiral coil is aligned with the axis of the pipeline.

[0013] The pipe is made of polymer material and carries liquid inside.

[0014] The transceiver terminal is located on the ground, uses broadband signal transmission, and has full-duplex transmission and reception capabilities. The ratio of the broadband signal bandwidth to the center frequency is greater than 10%.

[0015] The full-duplex transceiver function and the broadband signal are designed to increase the amount of information in the transmission link and improve the stability of the system. Because the transceiver terminals have full-duplex transceiver functionality, in the pipeline monitoring system, any two adjacent transceiver terminals have opposite operating states at the same time. These operating states include transmitting and receiving. For every two adjacent transceiver terminals, if one terminal is in transmitting mode and outputs an excitation electrical signal to its corresponding transceiver coil, then the other terminal is in receiving mode and receives the induced electromagnetic signal from its corresponding transceiver coil.

[0016] The transceiver terminal includes a comb spectrum generator, which is used to generate broadband harmonic signals to increase the amount of information in a single signal transmission; the excitation electrical signal output by the transceiver terminal to the corresponding transceiver coil or the induced electromagnetic signal received from the corresponding transceiver coil are both the harmonic signals.

[0017] The transceiver terminal transmits or receives harmonic signals through corresponding transceiver coils. When the transceiver terminal receives a harmonic signal from its corresponding transceiver coil, it extracts information such as interference and loss caused by environmental changes in the signal transmission link and outputs it to the host computer. When the transceiver terminal outputs an excitation electrical signal to its corresponding transceiver coil, the excitation electrical signal is a broadband harmonic signal that is output simultaneously in a single operation.

[0018] As an optional embodiment of the present invention, the pipeline monitoring system includes two transceiver coils, a relay coil group, and two transceiver terminals; a relay coil group is provided between the two transceiver coils, and the two transceiver coils are electrically connected to the two transceiver terminals respectively. The two transceiver terminals operate in opposite states at the same time, with one transceiver terminal outputting an excitation electrical signal to its corresponding transceiver coil, and the other transceiver terminal receiving an induced electromagnetic signal from its corresponding transceiver coil; the operating states of the two transceiver terminals can be interchanged at different times.

[0019] In the host computer, the classification model is constructed based on the support vector machine model, and the accuracy of the classification model is used as the population fitness function of the genetic algorithm. The penalty coefficient C and kernel function parameter γ of the classification model are used as the genes of individuals in the genetic algorithm, and the model is trained using the genetic algorithm.

[0020] (II) A pipeline monitoring method based on magnetic induction coils

[0021] The pipeline monitoring method includes the following steps:

[0022] S1. Using a host computer, the functions of each transceiver terminal are set as either a transmitting terminal or a receiving terminal, so that the transmitting terminals and receiving terminals are cascaded one by one. The transmitting terminal outputs an excitation electrical signal to the corresponding transceiver coil, so that the transceiver coil generates a coupled and transmitted electromagnetic field distribution around the pipe. The receiving terminal receives the induced electromagnetic signal from the corresponding transceiver coil.

[0023] Step S1 specifically involves:

[0024] S1.1 At the initial moment, the host computer sends instructions to each transceiver terminal to set the working state of each transceiver terminal to either a transmitting terminal or a receiving terminal, and to make the transmitting terminals and receiving terminals cascaded one by one in an alternating manner, and set the initial moment to the current moment.

[0025] S1.2 At the current moment, the transmitting terminal and the receiving terminal are turned on. After one signal transmission process is completed, the host computer sends a working state switching command to each transceiver terminal to reverse the working state of each transceiver terminal, so that the transmitting terminal becomes the receiving terminal and the receiving terminal becomes the transmitting terminal.

[0026] Each signal transmission process is as follows: First, the transmitting terminal outputs an excitation electrical signal to the corresponding transceiver coil, causing the transceiver coil to generate a coupled and transmitted electromagnetic field distribution around the pipe. The excitation electrical signal is a broadband signal, and at the same time, all transceiver coils output the same excitation electrical signal. Next, the receiving terminal receives the induced electromagnetic signal at the location of the corresponding transceiver coil. The induced electromagnetic signal is a broadband signal and carries attenuation rate information in the transmission link. At the same time, the induced electromagnetic signals obtained by different receiving terminals differ depending on the transmission link.

[0027] S1.3 If the monitoring is not completed, return to step S1.2 after one time step; if the monitoring is completed, shut down the host computer and all transceiver terminals.

[0028] S2. Each receiving terminal extracts the spectral amplitude and phase information from the induced electromagnetic signal it receives and outputs it to the host computer. The host computer integrates the spectral amplitude and phase information received from all receiving terminals into fused feature parameters.

[0029] Step S2 specifically involves:

[0030] S2.1 Each receiving terminal extracts the spectral amplitude and phase information of the induced electromagnetic signal from the induced electromagnetic signal it receives, and outputs the spectral amplitude and phase information to the software radio of the host computer.

[0031] S2.2 The software radio receives the spectrum amplitude and phase information extracted by each receiving terminal respectively. The spectrum amplitude and phase information output by each receiving terminal constitute a set of feature parameters. The spectrum amplitude and phase information output by all receiving terminals are aggregated to form a feature parameter set. Then, the feature parameter set is output to the data processing unit of the host computer.

[0032] S2.3 The data processing unit performs post-processing on all feature parameter sets to obtain fused feature parameters. The post-processing process is as follows: abnormal data is deleted from the feature parameter set, the format and structure of the remaining feature parameters are standardized, and then a weighted average method or a weighted median method is used to fuse the standardized feature parameters to obtain fused feature parameters. Abnormal data refers to electromagnetic signal fluctuations in the environment caused by factors other than pipe leaks, including but not limited to electromagnetic signal fluctuations caused by ground vibrations or weather changes. Deleting abnormal data from the feature parameter set is a data cleaning process, and abnormal data can be judged using conventional methods such as setting feature parameter thresholds.

[0033] S3. The fused feature parameters are input into a pre-trained classification model. The classification model outputs pipeline monitoring results, which include: no pipeline leaks, and region labels corresponding to the location of leaks. The training process of the pre-trained classification model in step S3 is as follows:

[0034] Step 1) Divide the pipeline to be inspected into several areas at equal intervals along the pipeline route and encode all areas sequentially. Use the codes as area labels and set an "abnormal" label corresponding to the pipeline that has no leakage.

[0035] Step 2) Obtain the fusion feature parameters when a leak occurs in each area and the fusion feature parameters when the pipeline does not leak by means of experimental collection. Combine the fusion feature parameters when a leak occurs in the same area with the area label of that area to form a feature data pair, and combine the fusion feature parameters when the pipeline does not leak with the no-anomaly label to form a feature data pair. All feature data pairs constitute a feature dataset. Divide the feature dataset into a training dataset and a test dataset.

[0036] Step 3) Construct a classification model based on the support vector machine model, use the accuracy of the classification model as the population fitness function of the genetic algorithm, and use the penalty coefficient C and kernel function parameter γ of the classification model as the genes of individuals in the genetic algorithm. Use the genetic algorithm to train the classification model based on the training dataset. After training, the optimized classification model is obtained.

[0037] Step 4) Use the test dataset to evaluate the performance of the optimized classification model. If the performance evaluation result meets the preset performance threshold, the optimized classification model is used as the trained classification model. If the performance evaluation result does not meet the preset performance threshold, the optimized classification model is optimized again according to the process in Step 3).

[0038] The beneficial effects of this invention are as follows:

[0039] (1) This invention utilizes a ring-shaped spiral coil to realize a magnetic field distributed along the circumference of the pipe. The magnetic flux of the magnetic field is almost distributed within the ring surface of the ring-shaped spiral coil, which reduces the influence of the lossy liquid inside the pipe. It has practical value for monitoring the leakage of polymer pipes. Moreover, it has a simple structure and low cost, and has application value in the energy industry, agriculture and other fields.

[0040] (2) Based on the characteristics of amplitude, phase, frequency shift, etc. of the induced electromagnetic signal, the present invention can directly detect the leakage situation and location. This pipeline monitoring system has a low dependence on underground sensors, which can reduce deployment and maintenance costs.

[0041] (3) The present invention uses machine learning algorithms to analyze the characteristics of induced electromagnetic signals, extracts features under different leakage conditions, trains multiple sets of data, establishes a classification model of induced electromagnetic signals, and improves the accuracy of leakage detection.

[0042] (4) In this invention, the transceiver terminal uses broadband signal transmission to obtain more feature information. Furthermore, it employs a transceiver terminal with full-duplex electromagnetic signal transmission and reception capabilities, where each terminal can serve as a distributed station. By fusing and processing the induced electromagnetic signals acquired by multiple distributed stations, high-precision pipeline monitoring can be achieved. Attached Figure Description

[0043] Figure 1 This invention relates to a pipeline leak detection system based on a magnetic induction toroidal spiral coil.

[0044] Figure 2 This is a comparative schematic diagram showing the structure and magnetic field distribution of the coil in the pipeline monitoring system of Embodiment 1 and the comparative example of the present invention; Figure 2 (a) is a schematic diagram of the structure of a conventional toroidal coil in the comparative example of the present invention; Figure 2 (b) is a schematic diagram of the structure and magnetic field distribution of the annular spiral coil in Embodiments 1 to 2 of the present invention.

[0045] Figure 3 This is a schematic diagram of the electromagnetic simulation results of the pipeline monitoring structure in Embodiment 1 and the comparative example of the present invention.

[0046] Figure 4 The following are the S-parameter (scattering parameter) spectra of the four coils under different leakage conditions in Embodiment 1 of the present invention; Figure 4 (a) represents the S-parameters (S) of coil ①. 11 Spectrum diagram; Figure 4 (b) represents the S-parameters (S) of coil ②. 21 Spectrum diagram; Figure 4 (c) represents the S-parameters of coil ③ (S 31 Spectrum diagram; Figure 4 (d) represents the S-parameters of coil ④ (S 41 ) Spectrum diagram.

[0047] Figure 5 The transmission and reflection spectra of Embodiment 1 of the present invention are shown when the liquid content in the pipeline is different. Figure 5 (a) is the reflection spectrum; Figure 5 (b) is the transmission spectrum.

[0048] Figure 6 This is a diagram showing the experimental results of the pipeline monitoring structure in Embodiment 2 of the present invention.

[0049] Figure 7 This is a schematic diagram of the training process of the classification model in this invention.

[0050] Among them, 1. Transceiver coil; 2. Relay coil group; 3. Traditional toroidal coil; 4. Toroidal spiral coil. Detailed Implementation

[0051] Exemplary embodiments of the present invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present invention are shown in the drawings, it should be understood that the present invention may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0052] The principle of this invention is as follows: When a lossy leak (such as industrial wastewater) from a pipeline enters the underground environment, it causes a change in the electromagnetic energy attenuation rate, resulting in a change in the electromagnetic signal output by the transceiver coil 1 near the leak area to its corresponding receiving terminal, thus introducing a disturbance signal. Specifically, the disturbance signal is caused by changes in the dielectric constant and permeability of the underground environment due to the pipeline leak. For any two adjacent transceiver coils 1, when leaks occur at different locations between the two transceiver coils 1, the electromagnetic signal output by one of the transceiver coils 1 to its corresponding receiving terminal will have different characteristics. These characteristics include the amplitude, phase, and frequency shift of the electromagnetic signal, which can be used to directly detect the condition and location of the leak.

[0053] Furthermore, when a liquid with electromagnetic loss is transmitted in the pipeline, and when the liquid content in the pipeline is different, the electromagnetic signal output by the transceiver coil 1 to its corresponding receiving terminal has different characteristics.

[0054] As an optional embodiment of the present invention, the pipeline monitoring system includes two transceiver coils 1 and two transceiver terminals. The two transceiver coils 1 are electrically connected to the two transceiver terminals respectively; the two transceiver terminals operate in opposite states at the same time, with one terminal outputting an excitation electrical signal to its corresponding transceiver coil 1, and the other terminal receiving an induced electromagnetic signal from its corresponding transceiver coil 1; the operating states of the two transceiver terminals can be interchanged at different times. A relay coil group 2 can also be added between the two transceiver coils 1 for the coupling and transfer of electromagnetic energy. This embodiment, by placing the two transceiver coils 1 and the relay coil group 2 in a section of pipeline prone to leakage, enables real-time monitoring of local leakage in that section of pipeline.

[0055] Furthermore, the transceiver coil 1 and relay coil group 2 can be cascaded and expanded as needed to accommodate additional pipelines. The cascaded pipeline monitoring system includes at least two transceiver coils 1, at least two transceiver terminals, and a host computer. Each transceiver terminal is electrically connected to a corresponding transceiver coil 1. The transceiver terminal outputs an excitation electrical signal to or receives an induced electromagnetic signal from its corresponding transceiver coil 1. The transceiver coil 1 receives the excitation electrical signal from its corresponding transceiver terminal and then generates a coupled electromagnetic field distribution outside the pipeline, or outputs an induced electromagnetic signal at its location to its corresponding transceiver terminal. The magnetic field of the transceiver coil 1 is distributed circumferentially along the pipeline. Each transceiver terminal communicates with the host computer to input the information carried in the received induced electromagnetic signal. The host computer processes the received information using a pre-trained classification model to obtain the pipeline monitoring results.

[0056] Furthermore, the pipeline monitoring system also includes a relay coil group 2. A relay coil group is set between two adjacent transceiver coils 1. Each relay coil group 2 includes several relay coils, and the magnetic field of each relay coil is distributed along the circumference of the pipeline.

[0057] Specifically, all transceiver coils 1 and relay coils are mounted on the outside of the pipeline and arranged at equal intervals along the pipeline's direction (the direction of pipeline extension) to achieve long-distance, all-around monitoring. The first and last coils along the pipeline's direction are both transceiver coils 1.

[0058] Furthermore, by appropriately distributing the transceiver coil 1 and the relay coil, the optimal electromagnetic energy coupling distance can be achieved, thus realizing the goal of monitoring at the lowest cost.

[0059] Specifically, all transceiver coils 1 and relay coils employ a toroidal spiral coil 4. The toroidal spiral coil 4 has a toroidal structure, with its axis aligned with the axis of the pipe. The toroidal spiral coil is formed by connecting the two ends of a spiral coil, which is arranged spirally along its central axis, which is distributed circumferentially along the pipe. The electromagnetic energy of the toroidal spiral coil 4 is primarily distributed within the coil's toroidal surface. The toroidal spiral coil 4 (including the connections between transceiver coils 1 and relay coils, and between relay coils) can perform electromagnetic energy coupling even when the pipe is bent, and is also suitable for electromagnetic signal detection.

[0060] Specifically, the pipes are made of polymer material and carry liquid inside.

[0061] Specifically, the transceiver terminal uses broadband signal transmission and has full-duplex electromagnetic signal transmission and reception capabilities.

[0062] Wideband signal transmission refers to a signal bandwidth to center frequency ratio exceeding 10%. Full-duplex transceiver functionality and wideband signals aim to increase the amount of information in the transmission link and improve system stability. Specifically, because the transceiver terminals have full-duplex transceiver functionality, in the pipeline monitoring system, each pair of adjacent transceiver terminals has opposite operating states at the same time. One transceiver terminal acts as a transmitting terminal, outputting an excitation electrical signal to its corresponding transceiver coil 1, while the other transceiver terminal acts as a receiving terminal, receiving the induced electromagnetic signal from its corresponding transceiver coil 1.

[0063] Specifically, the host computer includes a software-defined radio and a data processing unit. The software-defined radio receives broadband harmonic signals, extracts feature parameters from the signals, and transmits these parameters to the data processing unit. The data processing unit is equipped with a pre-trained classification model, which analyzes the received information to obtain pipeline monitoring results. The classification model is built based on a support vector machine (SVM) model. The accuracy of the classification model is used as the population fitness function of the genetic algorithm, and the penalty coefficient C and kernel parameter γ are used as the genes of individuals in the genetic algorithm. The model is trained using a genetic algorithm.

[0064] Furthermore, the transceiver terminal includes a comb spectrum generator for generating a series of harmonic signals at intervals to increase the amount of information in a single signal transmission. When the transceiver terminal receives harmonic signals from its corresponding transceiver coil 1, it extracts information such as interference and loss caused by environmental changes in the signal transmission link and outputs it to the host computer. When the transceiver terminal outputs an excitation electrical signal to its corresponding transceiver coil 1, the excitation electrical signal is a broadband harmonic signal that is output simultaneously in a single transmission.

[0065] The pipeline monitoring method for the above-mentioned pipeline monitoring system provided by the present invention includes the following steps:

[0066] S1. Using a host computer, configure each transceiver terminal as either a transmitting terminal or a receiving terminal, so that the transmitting and receiving terminals are cascaded in an interleaved manner, i.e., adjacent transceivers in the pipeline monitoring system interleave signals; the transmitting terminal is used to output excitation electrical signals to its corresponding transceiver coil 1, and the receiving terminal is used to receive induced electromagnetic signals from its corresponding transceiver coil 1. Step S1 is specifically as follows:

[0067] S1.1 At the initial moment, the host computer sends instructions to each transceiver terminal to set the working state of each transceiver terminal to either a transmitting terminal or a receiving terminal, and to make the transmitting terminals and receiving terminals cascaded one by one in an alternating manner, and set the initial moment to the current moment.

[0068] S1.2 At the current moment, the transmitting terminal and the receiving terminal are turned on. After one signal transmission process is completed, the host computer sends a working state switching command to each transceiver terminal to reverse the working state of each transceiver terminal, so that the transmitting terminal becomes the receiving terminal and the receiving terminal becomes the transmitting terminal.

[0069] Each signal transmission process is as follows: First, the transmitting terminal outputs an excitation signal to the corresponding transceiver coil 1, causing the transceiver coil 1 to generate a coupled electromagnetic field distribution around the pipe. The excitation signal is a broadband signal, and at the same time, all transceiver coils 1 output the same excitation signal. Next, the receiving terminal receives the induced electromagnetic signal at the location of the corresponding transceiver coil 1. The induced electromagnetic signal is a broadband signal and carries attenuation rate information in the transmission link. At the same time, the induced electromagnetic signals obtained by different receiving terminals differ depending on the transmission link.

[0070] S1.3 If the monitoring is not completed, return to step S1.2 after one time step; if the monitoring is completed, shut down the host computer and all transceiver terminals.

[0071] S2. Each receiving terminal extracts the spectral amplitude and phase information from the induced electromagnetic signal it receives and outputs it to the host computer. The host computer integrates the spectral amplitude and phase information received from all receiving terminals into fused feature parameters. Step S2 is specifically as follows:

[0072] S2.1 Each receiving terminal extracts the spectral amplitude and phase information of the induced electromagnetic signal from the induced electromagnetic signal it receives, and outputs the spectral amplitude and phase information to the software radio of the host computer.

[0073] S2.2 The software radio receives the spectrum amplitude and phase information extracted by each receiving terminal. The spectrum amplitude and phase information output by each receiving terminal constitute a set of characteristic parameters. The spectrum amplitude and phase information output by all receiving terminals are combined to form a set of characteristic parameters. Then the set of characteristic parameters is output to the data processing unit of the host computer.

[0074] S2.3 The data processing unit performs post-processing on all feature parameter sets to obtain fused feature parameters. The post-processing process is as follows: abnormal data is deleted from the feature parameter set, the format and structure of the remaining feature parameters are unified, and then the weighted average method or weighted median method is used to fuse the feature parameters after unifying the format and structure to obtain fused feature parameters.

[0075] Abnormal data refers to electromagnetic signal fluctuations in the environment caused by factors other than pipe leaks, including but not limited to electromagnetic signal fluctuations caused by ground vibrations or weather changes. Removing abnormal data from the feature parameter set is a data cleaning process, and abnormal data can be identified using conventional methods such as setting feature parameter thresholds.

[0076] S3. Input the fused feature parameters into the pre-trained classification model. The classification model outputs the pipeline monitoring results, which include: no pipeline leakage and the area label corresponding to the location of the leakage.

[0077] In step S3, the training process of the pre-trained classification model is as follows:

[0078] Step 1) Divide the pipeline to be inspected into several areas at equal intervals along the pipeline route and encode all areas in sequence. Use the codes as area labels, that is, use the codes of each area as the area labels of each area, and set the no-abnormal label corresponding to the pipeline not leaking.

[0079] Step 2) Obtain the fusion feature parameters when a leak occurs in each area and the fusion feature parameters when the pipeline does not leak by means of experimental collection. Combine the fusion feature parameters when a leak occurs in the same area with the area label of that area to form a feature data pair, and combine the fusion feature parameters when the pipeline does not leak with the no-anomaly label to form a feature data pair. All feature data pairs constitute a feature dataset. Divide the feature dataset into a training dataset and a test dataset.

[0080] Step 3) Construct a classification model based on the support vector machine model. Use the accuracy of the classification model as the population fitness function of the genetic algorithm, and use the penalty coefficient C and kernel function parameter γ of the classification model as the genes of individuals in the genetic algorithm. Use the genetic algorithm to train the classification model based on the training dataset. When the genetic algorithm runs to the preset number of iterations or the fitness reaches the preset threshold, after training, select the optimal individual with the highest fitness from the current population. Use the penalty coefficient C and kernel function parameter γ of the optimal individual as the parameters in the trained classification model to obtain the optimized classification model.

[0081] Step 4) Use the test dataset to evaluate the performance of the optimized classification model. If the performance evaluation result meets the preset performance threshold, the optimized classification model is used as the trained classification model. If the performance evaluation result does not meet the preset performance threshold, the optimized classification model is optimized again according to the process in Step 3).

[0082] Specific embodiments and comparative examples of the present invention are as follows:

[0083] Comparative Example

[0084] This comparative example includes a transceiver coil 1, a relay coil group 2, and a ground transceiver terminal. The transceiver coil 1 and the transceiver terminal are connected by a wired cable.

[0085] Both the transceiver coil 1 and the repeater coil are conventional toroidal coils 3, coaxially wound around the outside of the pipe. For example... Figure 2 As shown in (a), the magnetic flux of the conventional loop coil 3 mainly passes through the inside of the loop. Along the transmission path, the electromagnetic energy may be attenuated due to the lossy liquid in the polymer pipe.

[0086] Example 1

[0087] Depend on Figure 1 As shown, this embodiment includes a transceiver coil 1, a relay coil group 2, and a transceiver terminal located on the ground. The transceiver coil 1 and the transceiver terminal are connected via a wired cable. The relay coil consumes almost no energy, and its optimal deployment in terms of quantity and distance ensures sufficient received signal strength to capture changes caused by leakage. Both the transceiver coil 1 and the relay coil are coaxially wound to the outside of the pipe using annular spiral coils 4. The structure of the annular spiral coil 4 is as follows... Figure 2 As shown in (b), the magnetic flux of the toroidal spiral coil 4 is almost distributed within its coil toroidal surface, reducing the influence of the lossy liquid and making it practically valuable for monitoring leaks in polymer pipelines.

[0088] The toroidal spiral coil 4 requires no additional maintenance or complex sensor clusters in underground environments, and can directly detect leaks from the received signals. When liquid leaks from a pipeline into the soil environment, the soil's dielectric constant and magnetic permeability are affected, especially in cases of complex leaks from industrial wastewater transport. This physical change in the energy transfer path introduces signal perturbation. This perturbation essentially stems from changes in the induction intensity between adjacent coils, and its degree is proportional to the amount of leakage.

[0089] In both the embodiments and comparative examples of the present invention, coil ① and coil ④ serve as transceiver coil 1, and coil ② and coil ③ constitute relay coil group 2. Coil ① is electrically connected to the transmitting terminal, and coil ④ is electrically connected to the receiving terminal. The signal transmission process is as follows: after coil ① receives the excitation electrical signal output by the transmitting terminal, it generates an excitation signal at coil ①. The excitation signal is relayed by coil ② and coil ③, and finally output to the receiving terminal at coil ④.

[0090] Electromagnetic simulations were performed on the pipeline monitoring structures in the above embodiments and comparative examples. In the electromagnetic simulation models: the pipeline material was non-destructive polycarbonate with a relative permittivity of 2.9; the liquid inside the pipeline was water with a conductivity of 1.59 S / m, representing industrial wastewater; the environment was dry soil with a relative permittivity of 2.44 and a loss tangent of 0.0014. The specific dimensions of the pipeline monitoring system are shown in the table below:

[0091]

[0092] The simulation results are plotted as transmission curves, as shown below. Figure 3 As shown. When coil ① is excited, the signal strength-frequency curve received from coil ② is S. 21 The signal strength-frequency curve received from coil ③ is S. 31 The signal strength-frequency curve received from coil ④ is S. 41 Therefore, it can be seen that the toroidal spiral coil 4 can transmit signals, while the signal attenuation between traditional toroidal coils 3 is severe.

[0093] Furthermore, based on the near-field coupling effect of the annular spiral coil 4, the leakage detection function of the pipeline monitoring structure (pipeline monitoring system based on annular spiral coil) in this embodiment is verified by analyzing the changes in the transmitted and received signals.

[0094] During the verification process, a liquid column with a radius of 2 mm and a height of 47 mm was used to simulate leakage. The location of the liquid column was divided into three regions: between coil ① and coil ②, between coil ② and coil ③, and between coil ③ and coil ④. The verification results are as follows. Figure 4 As shown.

[0095] Figure 4 This is a spectrum of S-parameters of the four spiral coils under different leakage conditions in an embodiment of the present invention (a pipeline monitoring system based on a toroidal spiral coil). There are four leakage conditions: no system leakage, leakage occurring between coil ① and coil ②, leakage between coil ② and coil ③, and leakage between coil ③ and coil ④.

[0096] in, Figure 4 (a) The S-parameters (S) of coil ① under different leakage conditions in this embodiment of the invention. 11 Spectrum diagram; Figure 4 (b) The S-parameters (S_s) of coil ② under different leakage conditions according to the embodiments of the present invention. 21 Spectrum diagram; Figure 4 (c) The S-parameters (S) of coil ③ under different leakage conditions in this embodiment of the invention. 31 Spectrum diagram; Figure 4 (d) shows the S-parameters (S_s) of coil ④ under different leakage conditions in this embodiment of the invention.41 ) Spectrum diagram.

[0097] When there is a leak in the pipeline, the S-parameters of the four spiral coils all show varying degrees of shift and attenuation compared to when there is no leak, especially the received signal strength. By post-processing the data on this shift and attenuation, the extent and location of the pipeline leak can be deduced.

[0098] In this embodiment, the electric field of the annular spiral coil 4 inevitably penetrates into the pipe and is affected by the liquid content in the pipe. Figure 5 As shown, when the liquid content in the pipeline varies, the peaks in the transmission and reflection spectra exhibit redshift and attenuation. This non-destructive testing method provides a new approach for pipeline flow monitoring.

[0099] The method for pipeline monitoring based on the pipeline monitoring structure (pipeline monitoring system based on a ring-shaped spiral coil) in this embodiment is as follows:

[0100] 1) The ground-based transmitting terminal sends an electrical signal through a wired cable to excite the underground transceiver coil 1 (coil ①). The transceiver coil 1 generates a coupled electromagnetic field distribution. The electromagnetic signal generated by the transceiver coil 1 is coupled to the other end of the transceiver coil 1 (coil ④) through a multi-stage relay coil (coil ② and coil ③). The other end of the transceiver coil 1, i.e. coil ④, transmits the induced electrical signal to the receiving terminal connected to it.

[0101] 2) The receiving terminal extracts the spectrum amplitude and phase information from the received electromagnetic signal, and transmits the extracted spectrum amplitude and phase information to the software radio in the host computer via LoRa communication using a wireless sensor; the software radio receives the spectrum amplitude and phase information to form the characteristic parameters of the induced electromagnetic signal, and transmits the characteristic parameters of the electromagnetic signal to the data processing unit of the host computer.

[0102] 3) The data processing unit inputs the characteristic parameters of the received electromagnetic signals into a pre-trained classification model. After processing the characteristic parameters of the electromagnetic signals, the classification model outputs the pipeline monitoring results. The pipeline monitoring results include four outcomes: no system leakage, leakage occurring between coil ① and coil ②, between coil ② and coil ③, and between coil ③ and coil ④.

[0103] like Figure 7As shown, the training process of the classification model is as follows: First, the location of the pipeline is divided into regions, and electromagnetic signals from different regions during leakage are collected. The spectral amplitude, phase, frequency shift characteristics, and region labels of the collected electromagnetic signals are used as feature datasets to determine the leakage situation. Training and test sets are then used for training the classification model and evaluating its performance, respectively. Next, a Support Vector Machine (SVM) model is used for classifying the leakage region labels. The accuracy of the SVM model is used as the population fitness function of the genetic algorithm. The genetic algorithm optimizes the penalty coefficient C and the γ function of the kernel function. The optimal SVM model is trained using the genetic algorithm based on the training set, and the performance of the optimal SVM model is evaluated using the test set.

[0104] Example 2

[0105] Two annular spiral coils 4 are used to surround the plastic water pipe. One annular spiral coil 4 receives the excitation electrical signal, and the other annular spiral coil 4 outputs the induced electromagnetic signal.

[0106] When leakage occurs between the two annular spiral coils 4, the S-parameter spectrum of the induced electromagnetic signal is collected and acquired. 21 The result is as follows Figure 6 As shown.

[0107] The above are merely embodiments of this application and are not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A magnetic induction coil based pipeline monitoring system for monitoring whether a pipeline under the ground is leaking and the location of the leak, characterized in that: The pipeline monitoring system comprises at least two transceiving coils (1), at least two transceiving terminals and a host computer; The transceiving terminals are electrically connected with the transceiving coils (1) one by one, the transceiving terminals are used for outputting excitation electric signals to the corresponding transceiving coils (1) or receiving induced electromagnetic signals from the corresponding transceiving coils (1), and the transceiving coils (1) are used for generating coupling transmission electromagnetic field distribution outside the pipeline after receiving the excitation electric signals from the corresponding transceiving terminals or outputting induced electromagnetic signals of the positions where the transceiving coils (1) are located to the corresponding transceiving terminals. The functions of the transceiving terminals are respectively set as transmitting terminals or receiving terminals, the transmitting terminals and the receiving terminals are staggered and cascaded one by one, the transmitting terminals output excitation electric signals to the corresponding transceiving coils (1), and the receiving terminals receive induced electromagnetic signals from the corresponding transceiving coils (1). All the transceiving coils (1) adopt annular spiral coils (4), the axis of the annular spiral coil (4) is aligned with the axis of the pipeline, the annular spiral coil (4) is formed by connecting the two ends of the spiral coil, the magnetic flux of the annular spiral coil (4) is mainly distributed in the coil annulus, and the magnetic field of the transceiving coil (1) is distributed along the circumference of the pipeline; the transceiving terminals are all connected with the host computer in communication to input the information carried in the induced electromagnetic signals received by the transceiving terminals to the host computer, and the host computer is used for processing the received information by using a pre-trained classification model to obtain a pipeline monitoring result. The excitation electric signals outputted by the transceiving terminals to the corresponding transceiving coils (1) or the induced electromagnetic signals received by the transceiving terminals from the corresponding transceiving coils (1) are all broadband harmonic signals.

2. The magnetic induction coil based pipe monitoring system of claim 1, wherein: The pipeline monitoring system further comprises a relay coil group (2), one relay coil group (2) is arranged between every two adjacent transceiving coils (1), each relay coil group (2) comprises a plurality of relay coils; all the transceiving coils (1) and the relay coils are sleeved on the outside of the pipeline and are arranged at equal intervals along the pipeline, the first coil and the last coil on the pipeline are both transceiving coils (1), and the relay coils adopt annular spiral coils (4), the axis of the annular spiral coil (4) is aligned with the axis of the pipeline.

3. A magnetic induction coil based pipe monitoring system according to any of claims 1 or 2, characterized in that: The pipeline adopts a polymer material, and a liquid is transmitted in the pipeline.

4. The magnetic induction coil based pipe monitoring system of claim 1, wherein: The transceiving terminals are arranged on the ground, adopt broadband signal transmission, and have a duplex transceiving function, and the ratio of the signal bandwidth to the center frequency of the broadband signal is greater than 10%.

5. The magnetic induction coil based pipe monitoring system of claim 4, wherein: The pipeline monitoring system comprises two transceiving coils (1), one relay coil group (2) and two transceiving terminals, one relay coil group (2) is arranged between the two transceiving coils (1), the two transceiving coils (1) are electrically connected with the two transceiving terminals respectively, and the working states of the two transceiving terminals at the same time are opposite, one transceiving terminal outputs excitation electric signals to the corresponding transceiving coil (1), and the other transceiving terminal receives induced electromagnetic signals from the corresponding transceiving coil (1).

6. A magnetic induction coil based pipe monitoring system as claimed in claim 1, wherein: The transceiving terminal comprises a comb spectrum generator, and the comb spectrum generator is used for generating a harmonic signal.

7. The magnetic induction coil based pipe monitoring system of claim 1, wherein: The classification model is constructed based on a support vector machine model, and the accuracy of the classification model is used as the population fitness function of the genetic algorithm, and the penalty coefficient C and the kernel function parameter gamma of the classification model are used as the genes of the individuals in the genetic algorithm, and the genetic algorithm is used for training.

8. A method for monitoring a pipeline using the magnetic induction coil-based pipeline monitoring system according to any one of claims 1 to 7, characterized in that: The method comprises the following steps: S1, using the host computer to set the functions of each transceiver terminal as a transmitting terminal or a receiving terminal respectively, so that the transmitting terminals and the receiving terminals are staggered and cascaded one by one, the transmitting terminal outputs an excitation electrical signal to the corresponding transceiving coil (1), so that the transceiving coil (1) generates a coupled transmission electromagnetic field distribution outside the pipeline, and the receiving terminal receives an induced electromagnetic signal from the corresponding transceiving coil (1); S2, each receiving terminal extracts the frequency spectrum amplitude and phase information from the induced electromagnetic signal received by itself and outputs the information to the host computer, and the host computer integrates the frequency spectrum amplitude and phase information received from all receiving terminals into fusion feature parameters; S3, inputting the fusion feature parameters into a pre-trained classification model, and the classification model outputs a pipeline monitoring result, and the pipeline monitoring result includes: pipeline non-leakage and a region label corresponding to the leakage position.

9. The method of pipeline monitoring according to claim 8, wherein: The step S2 is specifically: S2.1, each receiving terminal extracts the frequency spectrum amplitude and phase information from the induced electromagnetic signal and outputs the information to the software radio of the host computer; S2.2, the software radio receives the frequency spectrum amplitude and phase information extracted by each receiving terminal respectively, the frequency spectrum amplitude and phase information output by each receiving terminal constitutes a group of feature parameters, all the frequency spectrum amplitude and phase information output by the receiving terminals are integrated to form a feature parameter set, and then the feature parameter set is output to the data processing unit of the host computer; S2.3, the data processing unit post-processes all the feature parameter sets to obtain fusion feature parameters; the post-processing process is: deleting abnormal data from the feature parameter set, unifying the format and structure of the remaining feature parameters, and then using a weighted average method or a weighted median method to fuse the feature parameters with unified format and structure to obtain the fusion feature parameters.

10. The method of claim 8, wherein: In the step S3, the training process of the pre-trained classification model is specifically: 1) dividing the pipeline to be detected into a plurality of regions along the pipeline direction at equal distances and sequentially coding all the regions, taking the code as a region label, and setting a non-abnormal label corresponding to pipeline non-leakage; 2) acquiring the fusion feature parameters when each region leaks and the fusion feature parameters when the pipeline does not leak through experimental acquisition, combining the fusion feature parameters when the same region leaks and the region label of the region to form a feature data pair, and combining the fusion feature parameters when the pipeline does not leak and the non-abnormal label to form a feature data pair, all the feature data pairs forming a feature data set, and dividing the feature data set into a training data set and a test data set; 3) constructing a classification model based on a support vector machine model, taking an accuracy rate of the classification model as a population fitness function of a genetic algorithm, and taking a penalty coefficient C and a kernel function parameter γ of the classification model as genes of individuals in the genetic algorithm, training the classification model according to the training data set by using the genetic algorithm, and obtaining an optimized classification model after the training is completed; 4) performing performance evaluation on the optimized classification model by using a test data set, if a performance evaluation result meets a preset evaluation performance threshold, taking the optimized classification model as a trained classification model, and if the performance evaluation result does not meet the preset evaluation performance threshold, optimizing the optimized classification model again according to a process in step 3).

Images