A pipeline leakage monitoring method and system based on multi-parameter collaborative discrimination

Abstract

The application provides a pipeline leakage monitoring method and system based on multi-parameter collaborative discrimination, and belongs to the technical field of pipeline monitoring; the method comprises the following steps: collecting multi-dimensional data by arranging a plurality of monitoring terminals along the target pipeline in an axial interval; correcting the residual error of the multi-dimensional data based on a coupling compensation model, and updating the compensation coefficient in real time through a coefficient self-adaptive adjustment mechanism; sending the corrected multi-dimensional data to a data processing center for data analysis based on a hierarchical data transmission strategy to obtain a fusion feature vector, and judging pipeline leakage based on a BP neural network; if leakage occurs, starting an audible and visual early warning and positioning the pipeline leakage point by using a collaborative positioning method; the application can realize effective monitoring of pipeline leakage and accurately position the leakage point when actual leakage occurs.

Description

Technical Field

[0001] This invention belongs to the field of pipeline monitoring technology, and in particular relates to a pipeline leakage monitoring method and system based on multi-parameter collaborative discrimination. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] Pipeline transportation, as a core mode of transporting fluids such as oil, natural gas, and chemical raw materials, boasts significant advantages including large capacity, low cost, and minimal environmental impact, playing a crucial role in actual production activities. However, due to multiple factors, pipelines pose a risk of leakage during long-term service, seriously threatening public safety and the environment. Therefore, the development of efficient and reliable pipeline leakage monitoring technologies is an urgent practical need.

[0004] Traditional pipeline leak monitoring technologies primarily rely on the capture and analysis of single physical parameters. These single-parameter monitoring methods suffer from inherent limitations such as poor universality, susceptibility to environmental interference, and low reliability, leading to unsatisfactory accuracy in pipeline leak monitoring. To overcome the shortcomings of single parameters, multi-parameter collaborative monitoring has become an inevitable trend in technological development. However, current technologies largely remain at the level of simply collecting and superimposing data from multiple sensors such as temperature, vibration, and pressure, thus generally exhibiting the following technical shortcomings: (1) It is impossible to delve into the inherent physical relationship and coupling mechanism between different parameters. Although this "simple superposition" method can improve the accuracy of pipeline leakage monitoring to a certain extent, it still has obvious shortcomings in sensitivity, accuracy and anti-interference ability under special working conditions such as complex environment and small leakage, so there are false alarms and missed alarms.

[0005] (2) Most existing methods improve monitoring accuracy through this combination of data monitoring. However, when a pipeline leak is detected, it can only alert relevant personnel to repair it by issuing an alarm, but cannot provide the specific location of the leak. Relevant personnel need to find the location themselves, which leads to a long time interval between the occurrence of the pipeline leak and the repair. Summary of the Invention

[0006] To overcome the shortcomings of the prior art, the present invention provides a pipeline leakage monitoring method and system based on multi-parameter collaborative discrimination, which can effectively monitor pipeline leakage and accurately locate the leakage point when a leakage actually occurs.

[0007] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions: The first aspect of this invention provides a pipeline leakage monitoring method based on multi-parameter collaborative discrimination.

[0008] A pipeline leakage monitoring method based on multi-parameter collaborative discrimination includes: By deploying multiple monitoring terminals at intervals along the axial direction of the target pipeline, multidimensional data of the outer wall of the pipeline can be acquired synchronously. A coupled compensation model is constructed to correct the residuals of the obtained multidimensional data, and the compensation coefficients are updated in real time by introducing an adaptive coefficient adjustment mechanism. Based on a hierarchical data transmission strategy, the corrected multidimensional data is sent to the data processing center for data analysis. Specifically, based on the cross power spectral density and phase spectrum of vibration signals from adjacent monitoring terminals in the multidimensional data, frequency bands with continuous and stable derivative values ​​and fluctuations less than a preset threshold are selected as effective leakage signal frequency bands. Based on the selection results, signal matching is performed and fusion feature vectors are extracted. Based on the fused feature vector, a pipeline leak is determined using a BP neural network; if a leak is determined, an audible and visual warning is activated and a collaborative positioning method is used to locate the pipeline leak point.

[0009] Furthermore, the coupling compensation model is expressed as: ; in, This indicates the temperature value after compensation. This represents the original temperature reading. Indicates ambient temperature. Indicates the peak value of vibration acceleration. Indicates the dominant frequency of the vibration signal. Indicates the relative humidity of the environment; This represents the first compensation coefficient. This represents the second compensation coefficient. This represents the third compensation coefficient.

[0010] Furthermore, the compensation coefficients are updated in real time by introducing an adaptive adjustment mechanism, including: calculating the compensation residual in real time based on the compensated temperature value; and dynamically updating the compensation coefficients, including the first compensation coefficient, the second compensation coefficient, and the third compensation coefficient, by using the gradient descent method when the compensation residual exceeds the preset residual threshold.

[0011] Furthermore, the hierarchical data transmission strategy is implemented based on a three-level transmission architecture. Specifically: terminal-level transmission is achieved by building a LORA short-range communication module in each monitoring terminal and establishing a self-organizing network between adjacent monitoring terminals; a relay node is set up for every preset number of monitoring terminals and integrates dual communication modules to achieve long-distance relay-level transmission; and center-level transmission between the relay node and the data processing center is achieved through a redundant transmission method.

[0012] Furthermore, when the corrected multidimensional data is sent to the data processing center for data analysis, the multidimensional data is encrypted with AES-256 and verified with CRC32.

[0013] Furthermore, based on the selection results, signal matching and fusion feature vector extraction are performed, including: based on the timestamp when acquiring multidimensional data, synchronizing the temperature data of multidimensional data with the selected vibration data so that the temperature data and vibration data at the same time correspond; subsequently, extracting time-domain features from the temperature data of the corresponding multidimensional data of adjacent monitoring terminals, and extracting time-domain and frequency-domain features from the selected vibration data to construct a multi-parameter feature vector as the fusion feature vector.

[0014] Furthermore, a collaborative positioning method is used to locate the pipeline leak point, including: first, using an FIR filter to perform bandpass filtering on the non-dispersion frequency band of the vibration signal, and calculating the signal delay between adjacent monitoring terminals through cross-correlation function; then, calculating the actual propagation speed based on the sound velocity of the fluid in the pipe in the free field, combined with the multidimensional factors affecting sound propagation during leakage; finally, determining the location of the pipeline leak point based on the actual propagation speed and the pipeline distance between adjacent monitoring terminals.

[0015] A second aspect of the present invention provides a pipeline leakage monitoring system based on multi-parameter collaborative discrimination.

[0016] A pipeline leakage monitoring system based on multi-parameter collaborative discrimination includes: multiple monitoring terminals, a data transmission module, a data processing center, an early warning module, and a pipeline leakage location module; The monitoring terminal is arranged at intervals along the axial direction of the target pipeline and includes an integrated sensor probe, a magnetic mounting base, a data acquisition unit, a power supply unit, and an adaptive compensation sensor calibration module. The data acquisition unit is used to synchronously acquire multidimensional data of the pipeline's outer wall. The adaptive compensation sensor calibration module includes an environmental temperature and humidity sensor and a vibration coupling compensation algorithm module. The vibration coupling compensation algorithm module integrates a coupling compensation model and an adaptive coefficient adjustment mechanism to perform residual correction on the acquired multidimensional data and update the compensation coefficients in real time. The data transmission module adopts a three-level transmission architecture, which is used to send the corrected multidimensional data to the data processing center for data analysis based on a hierarchical data transmission strategy. The data processing center includes a data receiving unit, a data preprocessing unit, a feature extraction unit, a multi-parameter fusion and discrimination unit, and a database unit. The data processing center is used to select frequency bands with continuously stable derivative values ​​and fluctuations less than a preset threshold as effective leakage signal frequency bands based on the cross-power spectral density and phase spectrum of vibration signals from adjacent monitoring terminals in multi-dimensional data; perform signal matching and extract fused feature vectors based on the selection results; and determine pipeline leakage based on a BP neural network according to the fused feature vectors. The early warning module includes an audible and visual alarm unit and an information push unit; if a leak is detected, the audible and visual alarm unit will activate the audible and visual warning and send the leak information through the information push unit. The pipeline leak location module is used to locate the pipeline leak point using a collaborative location method.

[0017] A third aspect of the present invention provides a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the steps of a pipeline leakage monitoring method based on multi-parameter collaborative discrimination as described in the first aspect of the present invention.

[0018] The fourth aspect of the present invention provides an electronic device, including a memory, a processor, and a program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of a pipeline leakage monitoring method based on multi-parameter collaborative discrimination as described in the first aspect of the present invention.

[0019] The above one or more technical solutions have the following beneficial effects: (1) This invention constructs a coupled compensation model that integrates multi-dimensional data and introduces a coefficient adaptive adjustment mechanism based on gradient descent, thereby achieving dynamic and accurate correction of the collected data and effectively separating the coupling effect between environmental interference and mechanical vibration. On this basis, this invention adopts an improved BP neural network model optimized for leakage characteristics, which deeply integrates and intelligently distinguishes the compensated high-precision temperature time-domain features (mean, rate of change, variance) with the non-dispersion frequency band time-frequency domain features (peak value, effective value, spectral centroid) of the vibration signal. This collaborative mechanism from data source compensation to intelligent fusion at the feature layer breaks through the simple superposition mode of existing multi-parameter monitoring, significantly improving the monitoring sensitivity, discrimination accuracy, and overall anti-interference capability of the system in complex environments such as high temperature, high humidity, and strong vibration, as well as under micro-leakage conditions, fundamentally reducing the probability of false alarms and missed alarms.

[0020] (2) When a pipeline leak is detected, this invention can automatically perform cross-correlation analysis on the non-dispersion frequency band of the screened vibration signal and accurately extract the propagation delay of the signal from adjacent monitoring terminals. Simultaneously, it employs a leak sound propagation velocity model that considers the pipe wall-fluid coupling effect, combined with the pipe material and geometric parameters, to calculate the signal propagation velocity in the actual pipe body in real time. Finally, based on the time delay and velocity, the precise distance between the leak point and the monitoring terminal is directly calculated using a positioning formula, and the actual pipeline mileage can be marked on a GIS map. Therefore, this invention can provide accurate leak location information, greatly shortening the time required for personnel to investigate and locate the leak, providing crucial support for rapid emergency response and repair decisions, and effectively controlling the spread and losses of leak accidents.

[0021] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0022] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0023] Figure 1 This is a flowchart of a pipeline leakage monitoring method based on multi-parameter collaborative discrimination in Embodiment 1 of the present invention. Detailed Implementation

[0024] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0025] It should be noted that the terminology used herein is for the purpose of describing particular implementations only and is not intended to limit the exemplary implementations of the present invention.

[0026] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0027] Example 1 This embodiment discloses a pipeline leakage monitoring method based on multi-parameter collaborative discrimination.

[0028] like Figure 1 As shown, a pipeline leakage monitoring method based on multi-parameter collaborative discrimination includes: Step S1: Simultaneously acquire multi-dimensional data of the outer wall of the pipeline by deploying multiple monitoring terminals at intervals along the axial direction of the target pipeline. Step S2: Construct a coupled compensation model, perform residual correction on the obtained multidimensional data, and update the compensation coefficients in real time by introducing a coefficient adaptive adjustment mechanism. Step S3: Based on the hierarchical data transmission strategy, the corrected multidimensional data is sent to the data processing center for data analysis, namely: based on the cross power spectral density and phase spectrum of the vibration signals of adjacent monitoring terminals in the multidimensional data, the frequency band with continuous and stable derivative values ​​and fluctuations less than the preset threshold is selected as the effective leakage signal frequency band; based on the selection results, signal matching is performed and fusion feature vectors are extracted. Step S4: Based on the fused feature vector, determine pipeline leakage using a BP neural network; if leakage is determined, activate audible and visual warnings and use a collaborative positioning method to locate the pipeline leak point.

[0029] Based on the above process, this invention can effectively monitor pipeline leaks and accurately locate the leak point when a leak actually occurs. To facilitate understanding of the technical solution of this invention, the specific implementation methods of this invention will be further explained and described below.

[0030] In step S1, multiple monitoring terminals are deployed at intervals along the axial direction of the target pipeline to synchronously acquire multidimensional data of the outer wall of the pipeline.

[0031] Monitoring terminals are arranged at intervals along the pipeline axis. The spacing between two adjacent monitoring terminals is determined based on the type of medium transported by the pipeline and the pipeline diameter. For example, for pipelines carrying flammable and explosive media such as oil and natural gas, the spacing is set to 50-100 meters; for pipelines carrying ordinary water or chemical raw materials, the spacing is set to 100-200 meters to ensure comprehensive monitoring coverage. In this embodiment, one monitoring terminal is arranged every 80 meters along the pipeline axis, for a total of 20 terminals. Based on the deployed monitoring terminals, the collected multidimensional data types and corresponding information are shown in Table 1.

[0032] Table 1. List of Multidimensional Data Types and Corresponding Information

[0033] In step S2, a coupled compensation model is constructed to correct the residuals of the obtained multidimensional data, and the compensation coefficients are updated in real time by introducing an adaptive coefficient adjustment mechanism.

[0034] Based on the characteristics of ambient temperature and humidity, vibration acceleration amplitude, and vibration frequency, a multi-dimensional coupled compensation model is constructed to achieve accurate correction of temperature acquisition data through dynamic coefficient adjustment. The coupled compensation model is expressed as follows: ; in, This indicates the temperature value after compensation. This represents the original temperature reading. Indicates ambient temperature. Indicates the peak value of vibration acceleration. Indicates the dominant frequency of the vibration signal. Indicates the relative humidity of the environment; , and These represent the first, second, and third compensation coefficients, calibrated experimentally. In this embodiment, , and Based on experience, the actual values ​​are as follows: , and .

[0035] Based on the compensated temperature value Real-time calculation of compensation residuals: ; in, Indicates compensation for residuals; This represents the standard temperature reference value, provided by a built-in high-precision calibration chip. When the compensation residual exceeds a preset residual threshold... At that time, the compensation coefficients, including the first, second, and third compensation coefficients, are dynamically updated using the gradient descent method. Specifically, when... hour: ; ; ; in, This represents the updated first compensation coefficient. This represents the updated second compensation coefficient. This represents the updated third compensation coefficient; This represents the learning rate, which is set to 0.001 in this embodiment to ensure stability during coefficient adjustment. Based on this, the initial coefficient values ​​are reset and the process iterates again every 10 minutes to avoid accumulating errors.

[0036] In step S3, the corrected multidimensional data is sent to the data processing center for data analysis based on a hierarchical data transmission strategy. Specifically, based on the cross-power spectral density and phase spectrum of vibration signals from adjacent monitoring terminals in the multidimensional data, frequency bands with continuously stable derivative values ​​and fluctuations less than a preset threshold are selected as effective leakage signal frequency bands. Signal matching is then performed based on the selection results, and fusion feature vectors are extracted. This can be achieved through the following methods: Step S3-1: Implement a hierarchical data transmission strategy based on a three-level transmission architecture.

[0037] 1) Terminal-level transmission: By embedding a LORA short-range communication module in each monitoring terminal, a self-organizing network is established between adjacent monitoring terminals to achieve terminal-level transmission (communication distance ≤ 1km), realizing local mutual transmission and backup of monitoring data to avoid data loss at single terminals.

[0038] 2) Relay-level transmission: One relay node is set up every 5-8 monitoring terminals, and a LoRa gateway and a 5G / NB-IoT module are integrated in the relay node. After receiving and aggregating the data transmitted from the terminal level through the LoRa gateway, the data is transmitted over long distances (communication distance ≤10km) through the 5G / NB-IoT module, which is suitable for long-distance pipelines in the field.

[0039] 3) Central-level transmission: The relay node and the data processing center adopt a redundant transmission method of "5G as the main method and wired as the auxiliary method". Specifically, the field section is transmitted through 5G, and the factory section is transmitted through wired transmission such as Ethernet and RS485. The transmission rate is ≥1Mbps and the data transmission success rate is ≥99.9%.

[0040] By employing redundant transmission methods, center-level transmission between relay nodes and the data processing center is achieved.

[0041] When sending the corrected multidimensional data to the data processing center for data analysis, the AES-256 encryption algorithm is used to encrypt the transmitted data, and a CRC32 checksum is added to ensure the security and integrity of data transmission.

[0042] Step S3-2: After the data receiving unit of the data processing center receives the data, the data preprocessing unit processes the data.

[0043] First, adopt The criteria are as follows: Abnormal temperature and vibration data caused by sensor malfunctions or sudden interference are filtered out. Subsequently, a moving average method is used to reduce random noise interference and smooth the data, thereby improving data stability. It should be noted that these outlier filtering and data smoothing operations are existing technologies; therefore, this embodiment will not provide a detailed description of their further implementation methods.

[0044] For the smoothed vibration data, calculate the cross-power spectral density of the signals collected by adjacent monitoring terminals. And extract the phase spectrum Subsequently, the phase spectrum is differentiated: ; in, Indicates frequency resolution. Indicates the first Angular frequency of each data point. Preset threshold. Frequency band If a continuous interval exists ( And the difference between adjacent derivative values ​​within the interval If so, this interval is selected as the non-dispersion frequency band, and only the vibration signal of this frequency band is retained for subsequent feature extraction.

[0045] Based on this, the temperature data of the multidimensional data is synchronized and aligned with the filtered vibration data according to the timestamp when the multidimensional data is acquired, so that the temperature data and vibration data at the same time correspond. Subsequently, the feature extraction unit extracts time-domain features, including temperature mean, temperature change rate, and temperature variance, from the temperature data processed in the above steps. Time-domain features and frequency-domain features are extracted from the vibration data. The time-domain features include vibration peak value and vibration effective value, and the frequency-domain features are obtained by performing Fourier transform on the vibration data, including spectral peak value and spectral centroid.

[0046] In step S4, based on the fused feature vector, a pipeline leak is determined using a BP neural network; if a leak is determined, an audible and visual warning is activated and a collaborative localization method is used to locate the pipeline leak point.

[0047] The extracted temperature and vibration feature parameters are combined into a feature vector, which is then input into an improved BP neural network fusion discrimination model. This model is trained using a large amount of leakage and normal operation sample data to establish a mapping relationship between the feature parameters and the pipeline's operating status. The model performs inference calculations on the input feature vector and outputs the pipeline leakage judgment result and confidence level, with the confidence level ranging from 0 to 1. When the confidence level is greater than a set threshold of 0.85, the pipeline is judged to be leaking; otherwise, the pipeline is judged to be operating normally.

[0048] An improved BP neural network fusion discriminant model includes an input layer, hidden layers, and an output layer. Specifically, the input layer contains n nodes to receive preprocessed vectors. The hidden layer adopts a double-hidden-layer structure; the first layer is used to extract local features, such as the instantaneous rate of temperature change; the second layer is used to couple global features, such as the influence of temperature on vibration zero drift. Based on this, this invention uses Leaky-ReLU instead of the existing Sigmoid and Tanh activation functions; simultaneously, a dynamic pruning mechanism is introduced, that is, retaining more nodes in the early stage of training, and then automatically pruning redundant nodes according to the weight contribution during training to prevent overfitting. The output layer contains m nodes, corresponding to the calibrated temperature value and the calibrated vibration value, used to output the predicted value after nonlinear mapping.

[0049] If a pipeline leak is detected, an audible and visual warning is immediately activated via the early warning module, and the leak information is sent to the mobile terminal of the management personnel via the information push unit. At the same time, the data processing center stores the raw monitoring data, preprocessed data, characteristic parameters, and judgment results in the database unit to provide data support for subsequent leak cause analysis and system optimization.

[0050] After determining that a pipeline leak has occurred, a collaborative location method is used to pinpoint the leak point. Specifically: 1) Time Delay Extraction: For the non-dispersion frequency band of the vibration signal, an FIR filter is set for bandpass filtering. The inverse Fourier transform of the cross power spectral density after filtering is calculated to obtain the cross-correlation function. The time delay of the signals from adjacent monitoring terminals is extracted based on the peak value of the function. .

[0051] 2) Calculation of the propagation speed of leakage sound: ; in, This represents the sound velocity of the fluid inside the tube in a free field. Indicates the bulk modulus of a fluid. This indicates the Young's modulus of the pipe wall. Indicates the inner diameter of the pipe wall; This indicates the pipe wall thickness; this parameter value is obtained from the pipeline operation and maintenance log. This indicates the speed at which the sound from the leak propagates.

[0052] 3) Calculation of leak location: ; in, The distance between adjacent monitoring terminals; The distance from the leak point to the first monitoring terminal can be used to convert the distance into the actual pipeline mileage using a GIS map, thereby determining the specific location of the leak.

[0053] Example 2 This embodiment discloses a pipeline leakage monitoring system based on multi-parameter collaborative discrimination.

[0054] A pipeline leakage monitoring system based on multi-parameter collaborative discrimination includes: multiple monitoring terminals, a data transmission module, a data processing center, an early warning module, and a pipeline leakage location module; The monitoring terminal is arranged at intervals along the axial direction of the target pipeline and includes an integrated sensor probe, a magnetic mounting base, a data acquisition unit, a power supply unit, and an adaptive compensation sensor calibration module. The data acquisition unit is used to synchronously acquire multidimensional data of the pipeline's outer wall. The adaptive compensation sensor calibration module includes an environmental temperature and humidity sensor and a vibration coupling compensation algorithm module. The vibration coupling compensation algorithm module integrates a coupling compensation model and an adaptive coefficient adjustment mechanism to perform residual correction on the acquired multidimensional data and update the compensation coefficients in real time. The data transmission module adopts a three-level transmission architecture, which is used to send the corrected multidimensional data to the data processing center for data analysis based on a hierarchical data transmission strategy. The data processing center includes a data receiving unit, a data preprocessing unit, a feature extraction unit, a multi-parameter fusion and discrimination unit, and a database unit. The data processing center is used to select frequency bands with continuously stable derivative values ​​and fluctuations less than a preset threshold as effective leakage signal frequency bands based on the cross-power spectral density and phase spectrum of vibration signals from adjacent monitoring terminals in multi-dimensional data; perform signal matching and extract fused feature vectors based on the selection results; and determine pipeline leakage based on a BP neural network according to the fused feature vectors. The early warning module includes an audible and visual alarm unit and an information push unit; if a leak is detected, the audible and visual alarm unit will activate the audible and visual warning and send the leak information through the information push unit. The pipeline leak location module is used to locate the pipeline leak point using a collaborative location method.

[0055] Furthermore, the monitoring terminal adopts an integrated design, including an integrated sensor probe, a data acquisition unit, a magnetic mounting base, a power supply unit, and an adaptive compensation sensor calibration module.

[0056] 1) Integrated sensing probe: The platinum resistance temperature sensing element and the piezoelectric acceleration sensing element are encapsulated in the same stainless steel shell. The inner wall of the shell is filled with thermally conductive silicone to ensure that the temperature sensing element is in close contact with the outer wall of the pipe. The vibration sensing element is rigidly connected to the shell with a distance of ≤5mm between them to ensure the time synchronization of temperature and vibration data acquisition (synchronization error ≤1ms). 2) Magnetic mounting base: It uses neodymium iron boron strong magnets as the adsorption core and is wrapped with polytetrafluoroethylene anti-corrosion layer. No welding or drilling is required during installation. It can be directly adsorbed on the outer wall of the pipe. The adsorption force is ≥50N and it is suitable for pipes with a diameter ≥100mm. At the same time, it avoids damaging the original anti-corrosion layer and insulation layer of the pipe.

[0057] 3) Data acquisition unit: Dual-channel signal conditioning circuit, used to provide independent and synchronous signal amplification, filtering and analog-to-digital conversion channels for temperature sensor and vibration sensor; by using the same sampling clock, it ensures that the acquisition timing of the two parameters is completely aligned.

[0058] 4) Power supply unit: It adopts a combination of lithium battery and solar charging module for power supply. The battery capacity is ≥5000mAh and supports low power consumption mode (standby current ≤10μA) to meet the power supply needs of long-term unattended operation in the field.

[0059] 5) Adaptive compensation type sensor calibration module, connected to the data acquisition unit, including an environmental temperature and humidity sensor and a vibration coupling compensation algorithm module.

[0060] Ambient temperature and humidity sensor: Used to collect ambient temperature (measurement range -40℃-85℃, accuracy ±0.2℃) and relative humidity (measurement range 0-100% RH, accuracy ±2% RH) around the monitoring terminal, providing environmental compensation basis for the temperature sensor.

[0061] The vibration coupling compensation algorithm module constructs a multi-dimensional coupling compensation model based on environmental temperature and humidity, vibration acceleration amplitude, and vibration frequency characteristics. It achieves accurate correction of temperature acquisition data through dynamic coefficient adjustment. Simultaneously, an adaptive coefficient adjustment mechanism is introduced to update compensation parameters. This improved algorithm enhances temperature measurement accuracy to ±0.03℃ and improves anti-interference capability by over 50% under complex conditions such as high temperature, high humidity, and strong vibration, resolving the problem of insufficient compensation in multi-factor coupling scenarios found in existing algorithms.

[0062] Furthermore, the data transmission module adopts a hierarchical data transmission strategy of "terminal level - relay level - center level".

[0063] 1) Terminal-level transmission: By embedding a LORA short-range communication module in each monitoring terminal, a self-organizing network is established between adjacent monitoring terminals to achieve terminal-level transmission (communication distance ≤ 1km), realizing local mutual transmission and backup of monitoring data to avoid data loss at single terminals.

[0064] 2) Relay-level transmission: One relay node is set up every 5-8 monitoring terminals, and a LoRa gateway and a 5G / NB-IoT module are integrated in the relay node. After receiving and aggregating the data transmitted from the terminal level through the LoRa gateway, the data is transmitted over long distances (communication distance ≤10km) through the 5G / NB-IoT module, which is suitable for long-distance pipelines in the field.

[0065] 3) Central-level transmission: The relay node and the data processing center adopt a redundant transmission method of "5G as the main method and wired as the auxiliary method". Specifically, the field section is transmitted through 5G, and the factory section is transmitted through wired transmission such as Ethernet and RS485. The transmission rate is ≥1Mbps and the data transmission success rate is ≥99.9%.

[0066] By employing redundant transmission methods, center-level transmission between relay nodes and the data processing center is achieved.

[0067] When sending the corrected multidimensional data to the data processing center for data analysis, the AES-256 encryption algorithm is used to encrypt the transmitted data, and a CRC32 checksum is added to ensure the security and integrity of data transmission.

[0068] Furthermore, the data processing center includes a data receiving unit, a data preprocessing unit, a feature extraction unit, a multi-parameter fusion and discrimination unit, and a database unit.

[0069] 1) Data receiving unit, used to receive temperature data and vibration data transmitted by the data transmission module.

[0070] 2) Data preprocessing unit, used to perform outlier removal, data smoothing and synchronization alignment on the received data. Outlier removal adopts the 3σ criterion, data smoothing adopts the moving average method, and synchronization alignment is based on timestamp to achieve time synchronization between temperature data and vibration data.

[0071] 3) Feature extraction unit, used to extract time-domain features such as temperature mean, temperature change rate, and temperature variance from preprocessed temperature data, and to extract time-domain and frequency-domain features such as vibration peak value, vibration effective value, vibration spectrum peak value, and spectrum centroid from vibration data.

[0072] 3) Multi-parameter fusion discrimination unit: It adopts a fusion discrimination model based on an improved BP neural network, takes the extracted temperature and vibration feature parameters as input, and realizes the discrimination of pipeline leakage status through model training and inference.

[0073] The BP neural network introduces a momentum factor and an adaptive learning rate on the basis of the existing BP neural network. The momentum factor is set to 0.8-0.9, and the adaptive learning rate is dynamically adjusted according to the changes in the model training error. When the error increases, the learning rate is reduced, and when the error decreases, the learning rate is increased, so as to accelerate the convergence speed of the model and avoid getting trapped in local optima.

[0074] 4) Database unit, used to store raw monitoring data, preprocessed data, characteristic parameter data, and discrimination result data.

[0075] Furthermore, the early warning module is connected to the data processing center and includes an audible and visual early warning unit and an information push unit.

[0076] When the multi-parameter fusion discrimination unit determines that a pipeline leak has occurred, the audible and visual early warning unit immediately issues an audible and visual alarm signal. At the same time, the information push unit sends the leak information (including the leak location, the confidence level of the leak determination, etc.) to the mobile terminals of relevant management personnel via SMS, APP push, etc., to achieve timely early warning.

[0077] Example 3 The purpose of this embodiment is to provide a computer-readable storage medium.

[0078] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of a pipeline leakage monitoring method based on multi-parameter collaborative discrimination as described in Embodiment 1 of this disclosure.

[0079] Example 4 The purpose of this embodiment is to provide an electronic device.

[0080] An electronic device includes a memory, a processor, and a program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps in a pipeline leakage monitoring method based on multi-parameter collaborative discrimination as described in Embodiment 1 of this disclosure.

[0081] The steps and methods involved in the apparatuses of Embodiments 2, 3, and 4 above correspond to those in Embodiment 1. For specific implementation details, please refer to the relevant description section of Embodiment 1. The term "computer-readable storage medium" should be understood as a single medium or multiple media including one or more instruction sets; it should also be understood as including any medium capable of storing, encoding, or carrying an instruction set for execution by a processor and enabling the processor to perform any of the methods in this invention.

[0082] Those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computer devices. Optionally, they can be implemented using computer-executable program code, thereby allowing them to be stored in a storage device for execution by a computer device, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. The present invention is not limited to any particular combination of hardware and software.

[0083] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A pipeline leakage monitoring method based on multi-parameter collaborative discrimination, characterized in that, include: By deploying multiple monitoring terminals at intervals along the axial direction of the target pipeline, multidimensional data of the outer wall of the pipeline can be acquired synchronously. A coupled compensation model is constructed to correct the residuals of the obtained multidimensional data, and the compensation coefficients are updated in real time by introducing an adaptive coefficient adjustment mechanism. The corrected multidimensional data is sent to the data processing center for data analysis based on a hierarchical data transmission strategy. Based on the cross power spectral density and phase spectrum of vibration signals from adjacent monitoring terminals in multidimensional data, frequency bands with continuous and stable derivative values ​​and fluctuations less than a preset threshold are selected as effective leakage signal frequency bands; based on the selection results, signal matching is performed and fusion feature vectors are extracted. Based on the fused feature vector, a pipeline leak is determined using a BP neural network; if a leak is determined, an audible and visual warning is activated and a collaborative positioning method is used to locate the pipeline leak point.

2. The pipeline leakage monitoring method based on multi-parameter collaborative discrimination as described in claim 1, characterized in that, The coupling compensation model is expressed as follows: ； in, This indicates the temperature value after compensation. This represents the original temperature reading. Indicates ambient temperature. Indicates the peak value of vibration acceleration. Indicates the dominant frequency of the vibration signal. Indicates the relative humidity of the environment; This represents the first compensation coefficient. This represents the second compensation coefficient. This represents the third compensation coefficient.

3. A pipeline leakage monitoring method based on multi-parameter collaborative discrimination as described in any one of claims 1-2, characterized in that, The compensation coefficients are updated in real time by introducing an adaptive adjustment mechanism, including: calculating the compensation residual in real time based on the compensated temperature value; and dynamically updating the compensation coefficients, including the first compensation coefficient, the second compensation coefficient, and the third compensation coefficient, by using the gradient descent method when the compensation residual exceeds the preset residual threshold.

4. The pipeline leakage monitoring method based on multi-parameter collaborative discrimination as described in claim 1, characterized in that, The hierarchical data transmission strategy is implemented based on a three-level transmission architecture. Specifically: terminal-level transmission is achieved by building a LORA short-range communication module in each monitoring terminal and establishing a self-organizing network between adjacent monitoring terminals; a relay node is set up for every preset number of monitoring terminals and integrates dual communication modules to achieve long-distance relay-level transmission; and center-level transmission between the relay node and the data processing center is achieved through a redundant transmission method.

5. The pipeline leakage monitoring method based on multi-parameter collaborative discrimination as described in claim 1, characterized in that, When the corrected multidimensional data is sent to the data processing center for data analysis, the multidimensional data is encrypted with AES-256 and checked with CRC32.

6. The pipeline leakage monitoring method based on multi-parameter collaborative discrimination as described in claim 1, characterized in that, Based on the selection results, signal matching and fusion feature vector extraction are performed, including: aligning the temperature data of the multidimensional data with the selected vibration data based on the timestamp when acquiring the multidimensional data, so that the temperature data and vibration data at the same time correspond; subsequently, extracting time-domain features from the temperature data of the corresponding multidimensional data of adjacent monitoring terminals, and extracting time-domain and frequency-domain features from the selected vibration data, in order to construct a multi-parameter feature vector as the fusion feature vector.

7. The pipeline leakage monitoring method based on multi-parameter collaborative discrimination as described in claim 1, characterized in that, The collaborative localization method for locating pipeline leaks includes: First, using an FIR filter to bandpass filter the non-dispersion frequency band of the vibration signal, and calculating the signal delay between adjacent monitoring terminals using a cross-correlation function; then, calculating the actual propagation speed based on the sound velocity of the fluid in the pipe in a free field, combined with the multidimensional factors affecting sound propagation during leakage; finally, determining the location of the pipeline leak point based on the actual propagation speed and the pipeline distance between adjacent monitoring terminals.

8. A pipeline leakage monitoring system based on multi-parameter collaborative discrimination, characterized in that, include: Multiple monitoring terminals, data transmission modules, data processing centers, early warning modules, and pipeline leak location modules; The monitoring terminal is arranged at intervals along the axial direction of the target pipeline and includes an integrated sensor probe, a magnetic mounting base, a data acquisition unit, a power supply unit, and an adaptive compensation sensor calibration module. The data acquisition unit is used to synchronously acquire multidimensional data of the pipeline's outer wall. The adaptive compensation sensor calibration module includes an environmental temperature and humidity sensor and a vibration coupling compensation algorithm module. The vibration coupling compensation algorithm module integrates a coupling compensation model and an adaptive coefficient adjustment mechanism to perform residual correction on the acquired multidimensional data and update the compensation coefficients in real time. The data transmission module adopts a three-level transmission architecture, which is used to send the corrected multidimensional data to the data processing center for data analysis based on a hierarchical data transmission strategy. The data processing center includes a data receiving unit, a data preprocessing unit, a feature extraction unit, a multi-parameter fusion and discrimination unit, and a database unit. The data processing center is used to select frequency bands with continuously stable derivative values ​​and fluctuations less than a preset threshold as effective leakage signal frequency bands based on the cross-power spectral density and phase spectrum of vibration signals from adjacent monitoring terminals in multi-dimensional data; perform signal matching and extract fused feature vectors based on the selection results; and determine pipeline leakage based on a BP neural network according to the fused feature vectors. The early warning module includes an audible and visual alarm unit and an information push unit; if a leak is detected, the audible and visual alarm unit will activate the audible and visual warning and send the leak information through the information push unit. The pipeline leak location module is used to locate the pipeline leak point using a collaborative location method.

9. A computer-readable storage medium having a program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps in the pipeline leakage monitoring method based on multi-parameter collaborative discrimination as described in any one of claims 1-7.

10. An electronic device, comprising a memory, a processor, and a program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the pipeline leakage monitoring method based on multi-parameter collaborative discrimination as described in any one of claims 1-7.

Images