Pipeline leakage monitoring method and monitoring system based on unmanned aerial vehicle inspection

Abstract

The invention discloses a pipeline leakage monitoring method and system based on unmanned aerial vehicle routing inspection, and relates to the technical field of pipeline leakage monitoring, and the method comprises the steps: a leakage judgment module compares a pipeline leakage index LI value with a preset threshold value theta, and judges whether Legt is within a certain sampling period continuously; if so, triggering leakage early warning, and activating the positioning module; the leakage positioning module calculates the three-dimensional coordinates (x, y, z) and the influence radius Ra of a leakage point according to the pipeline leakage index LI value in combination with the GPS coordinates of the unmanned aerial vehicle, the sound wave arrival time difference delta t at the air pressure sudden change moment and the wind speed vector W. According to the method, the temperature anomaly coefficient K1, the air pressure fluctuation coefficient K2 and the voiceprint matching coefficient K3 are integrated through the pipeline leakage evaluation model, the pipeline leakage index LI is output, comprehensive quantitative evaluation of the pipeline leakage degree is achieved, tiny leakage and normal working condition fluctuation can be effectively distinguished, and the leakage judgment accuracy is improved.

Description

Technical Field

[0001] This invention relates to the field of pipeline leakage monitoring technology, and in particular to a method and system for monitoring pipeline leakage based on unmanned aerial vehicle (UAV) inspection. Background Technology

[0002] With the continuous rise in global energy demand, long-distance pipeline transportation networks in the oil and gas, chemical, and other fields are constantly expanding in depth. As a key infrastructure connecting energy production and consumption, the safe and stable operation of pipelines is directly related to the continuity of energy supply, the protection of the ecological environment, and the safety of public life and property. Especially in areas with complex terrain and harsh environments, such as mountains, deserts, and around water bodies, pipeline leaks caused by corrosion, aging, or third-party damage not only result in huge resource waste but may also trigger secondary disasters such as explosions and environmental pollution. Therefore, building an efficient, accurate, and real-time pipeline leak monitoring system has become one of the core requirements for industry development.

[0003] Currently, pipeline leak monitoring technology has evolved from traditional manual inspections to intelligent monitoring. Traditional methods rely on personnel carrying portable equipment for periodic checks. Limited by labor costs, inspection frequency, and accessibility in complex terrain, this approach struggles to achieve 24 / 7, full-coverage monitoring and suffers from significant delays in identifying minute leaks. Fixed monitoring systems, such as hydraulic model monitoring based on pressure and flow, can achieve continuous monitoring, but suffer from high installation and maintenance costs, numerous blind spots (e.g., in remote sections of long-distance pipelines), and insufficient sensitivity to external pipeline leaks. In recent years, drone technology, with its advantages of mobility, wide coverage, and adaptability to complex terrain, has gradually become an important means of pipeline inspection. Its onboard infrared thermal imagers, barometric pressure sensors, and acoustic sensors can collect multi-dimensional data around the pipeline, providing a new technological path for leak monitoring.

[0004] However, existing UAV-based pipeline leak monitoring technologies still have many problems that need to be solved: First, the ability to fuse multi-source data is insufficient. Most systems rely on data from a single type of sensor, such as relying solely on infrared thermal imaging or acoustic signals to determine leaks. This is easily affected by environmental interference, such as changes in lighting, background noise, and weather conditions, which can lead to false alarms or missed alarms. Second, the leak detection model lacks collaborative analysis of multiple parameters, making it difficult to quantify the degree of leakage and resulting in low differentiation between minor leaks and fluctuations in normal operating conditions. Third, the accuracy of leak location is greatly affected by environmental factors, such as wind speed and air pressure gradients, making it difficult to meet the precise location requirements for emergency repairs.

[0005] Therefore, it is essential to invent a monitoring method and system for pipeline leaks based on drone inspections to solve the above problems. Summary of the Invention

[0006] The purpose of this invention is to provide a method and system for monitoring pipeline leaks based on unmanned aerial vehicle (UAV) inspection, so as to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a pipeline leakage monitoring method based on unmanned aerial vehicle (UAV) inspection, comprising the following steps:

[0008] S1. The pipeline status data acquisition module uses a drone equipped with multiple sensors to inspect the pipeline along a preset route and collect multimodal raw data within a certain flight altitude range to form a pipeline status dataset; the pipeline status dataset includes infrared thermal imaging data, air pressure fluctuation data and acoustic signature data.

[0009] S2, the pipeline status data preprocessing module performs spatiotemporal alignment, noise filtering, and data standardization on the pipeline status dataset;

[0010] S3, the pipeline status data analysis module analyzes the preprocessed pipeline status dataset and calculates the temperature anomaly coefficient K1, air pressure fluctuation coefficient K2 and sound signature matching coefficient K3. The temperature anomaly coefficient K1, air pressure fluctuation coefficient K2 and sound signature matching coefficient K3 are input into the pipeline leakage assessment model and the pipeline leakage index LI is output.

[0011] S4. The leakage detection module compares the pipeline leakage index LI value with the preset threshold θ. When LI>θ continuously within a certain sampling period, a leakage warning is triggered and the positioning module is activated.

[0012] S5. The leak location module calculates the three-dimensional coordinates (x, y, z) and influence radius R of the leak point based on the pipeline leakage index LI value, combined with the UAV's GPS coordinates, the sound wave arrival time difference Δt at the moment of sudden air pressure change, and the wind speed vector W. a ;

[0013] S6. The report generation module will generate the three-dimensional coordinates of the leak point, the LI value, and the radius of influence R. a The repair suggestions are integrated into a structured report and uploaded to the monitoring center.

[0014] Preferably, the infrared thermal imaging data includes the maximum temperature difference between adjacent frames, the ambient reference temperature, and the effective area ratio of abnormally high-temperature regions in a single frame of infrared image; the air pressure fluctuation data includes the instantaneous air pressure value and the air pressure gradient change rate; and the voiceprint feature data includes the spectral amplitude.

[0015] Preferably, the temperature anomaly coefficient K1 is specifically:

[0016]

[0017] Where, ΔT maxThe maximum temperature difference between adjacent frames, T0 is the ambient reference temperature, and T crit Where A is the critical temperature for thermal deformation of the pipe material, A0 is the critical threshold for determining the proportion of the high-temperature region in the pipe, m is the area weight attenuation coefficient, and e is the natural constant. radio This represents the effective area percentage of abnormally high-temperature regions in a single frame of infrared image.

[0018] Preferably, the pressure fluctuation coefficient K2 is specifically:

[0019]

[0020] Where P(t) is the instantaneous air pressure value at second t, σ P(t) Let σt be the standard deviation of air pressure in second t, σ0 be the threshold standard deviation under normal operating conditions, and η be the standard deviation of air pressure in second t. P(t) η is the rate of change of the pressure gradient. max is the historical maximum gradient value, and tanh is the hyperbolic tangent function.

[0021] Preferably, the voiceprint matching coefficient K3 is specifically:

[0022]

[0023] Where f is the discrete frequency point of the spectrum analysis, S(f) is the spectral amplitude of the real-time acquired sound signal at frequency f, R(f) is the spectral amplitude of the preset pipeline leakage acoustic signature template at frequency f, f1 is the lower limit frequency of the feature band, f2 is the upper limit frequency of the feature band, L is the propagation distance, a is the distance attenuation constant, and e is the natural constant.

[0024] Preferably, the pipeline leakage assessment model is as follows:

[0025]

[0026] Where LI is the pipeline leakage index, K1 is the temperature anomaly coefficient, K2 is the air pressure fluctuation coefficient, K3 is the acoustic signature matching coefficient, α is the pressure attenuation factor, and P... thr Where β is the pressure threshold, λ is the acoustic attenuation factor, and A is the global risk factor. thr η is the acoustic threshold, η is the overall sensitivity, and e is the natural constant.

[0027] Preferably, the method for calculating the three-dimensional coordinates of the leakage point is as follows:

[0028]

[0029] Where x0 is the horizontal coordinate of the UAV's GPS, y0 is the horizontal coordinate of the UAV's GPS, and v x Let v be the linear velocity component of the object in the x-direction. yis the linear velocity component of the object in the y direction, W x is the angular velocity component of the object about the x axis, W y is the angular velocity component of the object about the y axis, H is the flight altitude, ρ is the air density, g is the acceleration due to gravity, P d is the air pressure at the leakage point, P0 is the ambient atmospheric pressure, and Δt is the time difference of the sound wave arrival at the moment of sudden air pressure change.

[0030] Preferably, the influence radius R a has the following calculation formula:

[0031] The influence radius R a has the following calculation formula:

[0032]

[0033] where, R0 is the reference influence radius, k R is the leakage index amplification coefficient, ω is the wind speed correction coefficient, ||W|| is the wind speed modulus, U gas is the initial gas injection velocity, and e is the natural constant.

[0034] Preferably, the repair suggestions include:

[0035] When LI > LI1, stop the pipeline transportation emergently;

[0036] When LI2 < LI ≤ LI1, strengthen the monitoring;

[0037] When LI3 < LI ≤ LI2, prompt the potential risks.

[0038] The pipeline leakage monitoring system based on UAV inspection specifically includes the following modules:

[0039] The pipeline status data acquisition module is used to inspect the pipeline along a preset route by a UAV equipped with multiple sensors, collect multi-modal raw data within a certain flight altitude range, and form a pipeline status data set; the pipeline status data set includes infrared thermal imaging data, air pressure fluctuation data, and voiceprint feature data;

[0040] The pipeline status data preprocessing module is used to perform spatio-temporal alignment, noise filtering, and data standardization processing on the pipeline status data set;

[0041] The pipeline status data analysis module is used to analyze the preprocessed pipeline status data set, calculate the temperature anomaly coefficient K1, air pressure fluctuation coefficient K2, and voiceprint matching coefficient K3, input the temperature anomaly coefficient K1, air pressure fluctuation coefficient K2, and voiceprint matching coefficient K3 into the pipeline leakage evaluation model, and output the pipeline leakage index LI;

[0042] The leakage detection module is used to compare the pipeline leakage index LI value with a preset threshold θ. When LI>θ continuously within a certain sampling period, a leakage warning is triggered and the location module is activated.

[0043] The leak location module is used to calculate the three-dimensional coordinates (x, y, z) and influence radius R of the leak point based on the pipeline leak index LI value, combined with the UAV's GPS coordinates, the sound wave arrival time difference Δt at the moment of sudden air pressure change, and the wind speed vector W. a ;

[0044] The report generation module is used to generate the three-dimensional coordinates of the leak point, the LI value, and the radius of influence R. a The repair suggestions are integrated into a structured report and uploaded to the monitoring center.

[0045] The technical effects and advantages of this invention are as follows:

[0046] 1. This invention utilizes a multi-sensor-equipped UAV to collect infrared thermal imaging data, air pressure fluctuation data, and acoustic signature data through a pipeline status data acquisition module, forming a multi-modal pipeline status dataset. This enables multi-dimensional monitoring of pipeline status, which, compared to single-sensor data, can more comprehensively capture leakage characteristics and reduce false alarms and missed alarms caused by environmental interference.

[0047] 2. This invention uses a leak location module to calculate the three-dimensional coordinates (x, y, z) of the leak point and its influence radius R based on the pipeline leakage index LI value, combined with the UAV's GPS coordinates, the sound wave arrival time difference Δt at the moment of sudden air pressure change, and the wind speed vector W. a By integrating multiple parameters, the accuracy of leak point location and the accuracy of impact range assessment are improved, effectively solving the problem that the location accuracy is greatly affected by environmental factors in the existing technology, and meeting the precise location requirements for emergency repairs;

[0048] 3. This invention integrates the temperature anomaly coefficient K1, air pressure fluctuation coefficient K2, and acoustic fingerprint matching coefficient K3 through a pipeline leakage assessment model, and outputs a pipeline leakage index LI, which realizes a comprehensive quantitative assessment of the degree of pipeline leakage. It can effectively distinguish between minor leaks and normal operating condition fluctuations, and improve the accuracy of leakage judgment.

[0049] 4. This invention classifies pipeline leaks into different levels based on the pipeline leakage index (LI) value through repair suggestions, thereby achieving differentiated treatment of pipeline leaks, improving the pertinence and efficiency of emergency response, and ensuring the safety of pipeline operation. Attached Figure Description

[0050] Figure 1 This is a schematic diagram of the method steps of the present invention.

[0051] Figure 2 This is a schematic diagram of the system module connections of the present invention. Detailed Implementation

[0052] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0053] This invention provides, for example Figure 1 The pipeline leak monitoring method based on drone inspection, as shown, includes the following steps:

[0054] S1. The pipeline status data acquisition module uses a drone equipped with multiple sensors to inspect the pipeline along a preset route and collect multimodal raw data within a certain flight altitude range to form a pipeline status dataset; the pipeline status dataset includes infrared thermal imaging data, air pressure fluctuation data and acoustic signature data.

[0055] Furthermore, in the above technical solution, the infrared thermal imaging data includes the maximum temperature difference between adjacent frames, the ambient reference temperature, and the effective area ratio of abnormally high-temperature regions in a single frame infrared image; the air pressure fluctuation data includes the instantaneous air pressure value and the air pressure gradient change rate; and the voiceprint feature data includes the spectral amplitude.

[0056] It should be noted that the aforementioned certain height range is from 5 meters to 50 meters;

[0057] The maximum temperature difference between adjacent frames is dynamically acquired by an infrared thermal imager mounted on a drone. The thermal imager takes thermal imaging sequences of the pipe surface at a fixed frequency along a preset flight path. After spatiotemporal alignment processing to ensure geographic coordinate consistency, the temperature difference of pixels at the same position in adjacent images is calculated frame by frame, and finally the maximum value of the temperature difference between all adjacent frames is extracted.

[0058] The system first captures infrared images of the pipe surface at a fixed frequency, such as 10 frames per second, using a thermal imager mounted on a drone. Then, it dynamically calibrates the temperature reference for each frame, extracts the average temperature of non-pipe areas, such as soil or vegetation, as the environmental reference temperature, and superimposes a preset temperature difference threshold, such as 15°C, to obtain the high temperature determination threshold. Finally, it generates a binary mask using an image segmentation algorithm, with high-temperature pixels set to 1 and the rest to 0. The ratio of the total number of high-temperature pixels to the total number of pixels in the image is calculated to obtain the effective area ratio of the abnormal high-temperature area in the single-frame infrared image.

[0059] The instantaneous air pressure value is obtained by sampling the air pressure data around the pipeline in real time at a fixed frequency of 10Hz using a high-precision air pressure sensor carried by the UAV, such as a MEMS piezoelectric barometer. Each sampling point is marked as P(t), which represents the instantaneous air pressure value at second t. The sensor needs to be equipped with a windproof cover and a low-pass filter to suppress rotor turbulence interference to ensure data reliability.

[0060] The spectral amplitude is obtained by capturing the original acoustic signal at a sampling rate of 48kHz using a high-sensitivity microphone array mounted on the UAV. After suppressing rotor noise with an adaptive filter, the frequency domain spectrum is generated by Fast Fourier Transform (FFT). Finally, the amplitude S(f) of all discrete frequency points f within the feature frequency band [f1, f2] is extracted. The discrete frequency resolution is determined by the number of FFT points; for example, a 4096-point FFT corresponds to a resolution of 11.7Hz.

[0061] The high-speed data processor for the rate of change of air pressure gradient performs numerical differentiation on the sequence of instantaneous air pressure values ​​P(t), and uses the central difference method to improve accuracy:

[0062]

[0063] Wherein, △t s The differential step size, Δt, is directly determined by the sensor sampling frequency, such as 10Hz. s =0.1s;

[0064] S2, the pipeline status data preprocessing module performs spatiotemporal alignment, noise filtering, and data standardization on the pipeline status dataset;

[0065] It is important to know that the spatiotemporal alignment is based on the high-precision timestamp of the UAV's GPS. The data collected by different sensors such as infrared thermal imagers, barometric pressure sensors, and microphone arrays are synchronized along the time axis to eliminate the time difference caused by the difference in sampling frequency. At the same time, combined with the UAV's real-time GPS coordinates and flight trajectory, the pixel position of infrared thermal imaging, barometric pressure sampling point, and acoustic fingerprint collection location are mapped to the same geographic coordinate system to ensure that the pipelines corresponding to different modal data are in the same spatial position.

[0066] The noise filtering employs targeted methods for different data types. Infrared thermal imaging data is smoothed by Gaussian or median filtering to suppress illumination changes and jitter noise. Barometric pressure data is filtered by wind shields and low-pass filters to reduce airflow and rotor turbulence interference. Acoustic data is filtered by adaptive filters to suppress low-frequency rotor noise and extracted by bandpass filters to remove environmental background noise.

[0067] The data standardization process involves normalizing or standardizing data of different dimensions, such as infrared temperature values, instantaneous air pressure values, and acoustic signature spectrum amplitudes, to unify them to the same scale and eliminate differences in magnitude. The preprocessed dataset, due to its characteristics of time synchronization, spatial consistency, noise suppression, and scale uniformity, provides a reliable foundation for subsequent feature analysis and leak determination.

[0068] S3, the pipeline status data analysis module analyzes the preprocessed pipeline status dataset and calculates the temperature anomaly coefficient K1, air pressure fluctuation coefficient K2 and sound signature matching coefficient K3. The temperature anomaly coefficient K1, air pressure fluctuation coefficient K2 and sound signature matching coefficient K3 are input into the pipeline leakage assessment model and the pipeline leakage index LI is output.

[0069] Furthermore, in the above technical solution, the temperature anomaly coefficient K1 is specifically defined as follows:

[0070]

[0071] Where, ΔT max The maximum temperature difference between adjacent frames, T0 is the ambient reference temperature, and T crit Where A is the critical temperature for thermal deformation of the pipe material, A0 is the critical threshold for determining the proportion of the high-temperature region in the pipe, m is the area weight attenuation coefficient, and e is the natural constant. radio This represents the effective area percentage of abnormally high-temperature regions in a single frame of infrared image.

[0072] It should be noted that the critical temperature for heat deformation of the pipe material depends on the pipe material, such as 300℃ for steel and 120℃ for PE plastic, which is directly retrieved from the engineering database;

[0073] The critical threshold A0 for the proportion of high-temperature area in the pipeline is determined comprehensively based on the actual engineering conditions of the specific pipeline. Specifically, engineers will combine pipeline type, such as oil or gas transmission, pipe material properties, such as steel or PE plastic, and industry safety standards. They will statistically analyze the proportion of high-temperature area in infrared images corresponding to historical leakage events and introduce safety margins for back-calculation. Finally, this critical threshold will be pre-stored in the engineering database for system calls. For example, A0 = 3% is set for oil pipelines, while A0 = 1.5% is set for gas pipelines with higher risks.

[0074] The default value of the area weight attenuation coefficient m is set to 1. In practice, it is calibrated according to the pipeline type, pipe material properties and environmental conditions. For example, for high-risk pipelines, such as gas pipelines, m is set higher, such as m>1, to quickly respond to small-area leaks; for low-risk pipelines, such as oil pipelines, m is set lower, such as m<1, to reduce false alarms; where m∈[0.5,2].

[0075] Furthermore, in the above technical solution, the pressure fluctuation coefficient K2 is specifically:

[0076]

[0077] Where P(t) is the instantaneous air pressure value at second t, σ P(t) Let σt be the standard deviation of air pressure in second t, σ0 be the threshold standard deviation under normal operating conditions, and η be the standard deviation of air pressure in second t. P(t) η is the rate of change of the pressure gradient. max is the historical maximum gradient value, and tanh is the hyperbolic tangent function.

[0078] It is important to know that the standard deviation of air pressure σ within the t-th second is... P(t) The following is calculated in real time by the embedded processor at the end of each second, based on all sampling points within that second:

[0079]

[0080] Where, μ P(t) P is the average value of the sampling points in that second. i Let N be the instantaneous air pressure value at the i-th sampling point within the t-th second, and N be the number of samples.

[0081] The normal operating condition standard deviation threshold σ0 is dynamically updated through a sliding time window mechanism, such as a 30-minute window.

[0082]

[0083] Where median is the median, IQR is the interquartile range, and Δt0 is the time step of the sliding window, which is set to 1 second by default.

[0084] Furthermore, in the above technical solution, the voiceprint matching coefficient K3 is specifically as follows:

[0085]

[0086] Where f is the discrete frequency point of the spectrum analysis, S(f) is the spectral amplitude of the real-time acquired sound signal at frequency f, R(f) is the spectral amplitude of the preset pipeline leakage acoustic signature template at frequency f, f1 is the lower limit frequency of the feature band, f2 is the upper limit frequency of the feature band, L is the propagation distance, a is the distance attenuation constant, and e is the natural constant.

[0087] It should be noted that the lower limit frequency f1 and the upper limit frequency f2 of the characteristic frequency band are determined by spectral analysis of laboratory simulated leaks or historical leak events. For example, for the high-frequency hissing sound of gas pipelines, f1 = 8kHz and f2 = 40kHz are set. These parameters are determined based on the type and material of the pipeline being inspected.

[0088] The preset pipeline leakage acoustic signature template uses a spectral amplitude R(f) at frequency f. This is achieved by precisely controlling valve opening and fluid pressure, such as 10MPa for gas pipelines and 4MPa for liquid pipelines, simulating leakage conditions with different orifice diameters. Simultaneously, a high-sensitivity microphone array is used to collect the original acoustic signal at a drone inspection altitude of 5-50 meters. After adaptive filtering to eliminate background noise, a spectrum is generated using Fast Fourier Transform (FFT), and characteristic frequency bands are extracted according to pipeline type: gas pipelines focus on high-frequency hissing sounds of 8-40kHz, and oil pipelines focus on low-frequency turbulent sounds of 100Hz-2kHz. The spectral amplitude S(f) within the characteristic frequency band [f1, f2] is normalized to eliminate the influence of volume differences. Finally, the spectral amplitude of all valid samples from the same type of pipeline is averaged to form a standardized acoustic signature template.

[0089] The propagation distance L is based on the sound wave arrival time difference positioning principle, and the spatially distributed microphone array carried by the UAV records the time difference Δt of the leaked sound wave arriving at different microphones in real time. c Combined with the ambient temperature T measured by the airborne temperature and humidity sensor c Used to calibrate the speed of sound c = 331 + 0.6 × T c Then, using the formula L=c×△t c Calculated;

[0090] The value of the distance attenuation constant 'a' is set between 0.02 and 0.05, and is automatically adjusted according to the pipeline type, such as 'a' = 0.03 for gas pipelines and 'a' = 0.04 for oil pipelines.

[0091] Furthermore, in the above technical solution, the pipeline leakage assessment model is specifically as follows:

[0092]

[0093] Where LI is the pipeline leakage index, K1 is the temperature anomaly coefficient, K2 is the air pressure fluctuation coefficient, K3 is the acoustic signature matching coefficient, α is the pressure attenuation factor, and P... thr Where β is the pressure threshold, λ is the acoustic attenuation factor, and A is the global risk factor. thr η is the acoustic threshold, η is the overall sensitivity, and e is the natural constant.

[0094] It is important to know that the value of the pressure attenuation factor α is set as follows: for high-pressure gas pipelines, α is set to a larger value, i.e., α≥1, to enhance sensitivity; for low-pressure oil or water pipelines, the value of α is appropriately reduced, i.e., α<1, to reduce false alarms; where α∈[0.5, 2.0], the actual setting needs to be optimized by gradient descent based on historical leakage data.

[0095] The pressure threshold P thrThe value is set as follows: for gas pipelines, it is set to 0.3 to 0.5; for oil pipelines, it is set to 0.6 to 0.8.

[0096] The acoustic attenuation factor β is set as follows: for gas pipelines, high-frequency sound waves attenuate quickly, so β ​​takes a larger value, i.e., β≥1, to compensate for signal loss; for oil pipelines, low-frequency sound waves have strong anti-interference capabilities, so β ​​takes a smaller value, i.e., β<1; where β∈[0.5, 3.0];

[0097] The global risk factor λ is set to 0.1–0.3 in normal areas to suppress false alarms; and to 0.7–1.0 in densely populated areas or ecological protection areas to enhance response.

[0098] The acoustic threshold A thr Value settings: For gas pipelines, set to 0.6–0.7; for oil pipelines, set to 0.75–0.85.

[0099] The default value of the overall sensitivity η is 1, which can be adjusted within the range of 0.5 to 2.0.

[0100] S4. The leakage detection module compares the pipeline leakage index LI value with the preset threshold θ. When LI>θ continuously within a certain sampling period, a leakage warning is triggered and the positioning module is activated.

[0101] It should be noted that the initial value of the preset threshold θ is set to 0.75, and later a new preset threshold θ is set based on 90% of the LI value of historical leakage events.

[0102] The sampling period can be set to be no less than 3 times, with 3 times as the default, but it can be adjusted.

[0103] S5. The leak location module calculates the three-dimensional coordinates (x, y, z) and influence radius R of the leak point based on the pipeline leakage index LI value, combined with the UAV's GPS coordinates, the sound wave arrival time difference Δt at the moment of sudden air pressure change, and the wind speed vector W. a ;

[0104] Furthermore, in the above technical solution, the method for calculating the three-dimensional coordinates of the leakage point is as follows:

[0105]

[0106] Where x0 is the horizontal coordinate of the UAV's GPS, y0 is the horizontal coordinate of the UAV's GPS, and v x Let v be the linear velocity component of the object in the x-direction. y W represents the linear velocity component of the object in the y-direction. x W represents the angular velocity component of the object about the x-axis. yLet H be the angular velocity component of the object about the y-axis, H be the flight altitude, ρ be the air density, g be the acceleration due to gravity, and P be the acceleration due to gravity. d P0 is the ambient atmospheric pressure, and Δt is the time difference of sound wave arrival at the moment of sudden change in air pressure.

[0107] It is important to know the linear velocity component v of the object in the x-direction. x and the linear velocity component v of the object in the y direction y It was obtained by differential calculation using real-time GNSS positioning data from the UAV;

[0108] The angular velocity component W of the object about the x-axis x and the angular velocity component W of the object about the y-axis y The angular rate of the fuselage's rotation about its axis is measured directly using an IMU;

[0109] The flight altitude H is obtained in real time by the GNSS receiver carried by the UAV, and its elevation value, i.e. the Z coordinate, is directly used as the flight altitude data.

[0110] The ambient atmospheric pressure P0 is obtained by hovering a drone over a safe area upstream of the inspection point or pipeline, and then continuously collecting air pressure data through a high-precision air pressure sensor, such as a MEMS piezoelectric barometer.

[0111] The gas pressure P at the leak point d By combining the sampled value P(t) of the pressure sensor at the moment of sudden change with historical data, it can be deduced that when the drone detects a sudden drop or rise in air pressure, the instantaneous air pressure value P(t) at that moment is the air pressure at the leak point. d ;

[0112] The time difference Δt of sound wave arrival at the moment of sudden change in air pressure is used to activate the spatially distributed microphone group at the bottom when the air pressure sensor triggers the sudden change alarm. The system accurately calculates the timestamps of sound wave arrival at each microphone, such as t1, t2, and t3, through a cross-correlation algorithm, and takes the maximum effective time difference between any two microphones, such as Δt = max(|t2-t1|, |t3-t1|).

[0113] Furthermore, in the above technical solution, the radius of influence R a The calculation formula is:

[0114]

[0115] Where R0 is the reference influence radius, k R ω is the leakage exponential amplification factor, ω is the wind speed correction factor, ||W|| is the wind speed modulus, and U gas Let be the initial velocity of the gas injection, and e be the natural constant.

[0116] It should be noted that the reference influence radius R0 is preset according to the pipeline type: if it is a gas transmission pipeline, R0 = 100m; if it is an oil transmission pipeline, R0 = 80m;

[0117] The leakage index amplification factor k R is set to 1.5 - 2.0, with a default setting of 1.8 and can be adjusted;

[0118] The value of the wind speed correction factor ω is set to 0.1 - 0.3 and can be adjusted according to the following environment: if it is a strong wind environment, i.e., ||W|| > 5m / s, take the upper limit, ω = 0.3; if it is a light wind environment, i.e., ||W|| < 2m / s, take the lower limit, ω = 0.1;

[0119] The wind speed modulus ||W|| is measured in real - time by an ultrasonic wind speed meter carried by a drone for the wind speed vector W=(U x , U y , U z ), and the calculation formula is

[0120] The initial gas injection velocity U gas The calculation formula is where ρ gas is the density of the medium in the pipeline. If it is a gas transmission pipeline, take the density of natural gas, ρ gas ≈0.8kg / m 3 ; if it is an oil transmission pipeline, take the density of crude oil, ρ gas ∈[800kg / m 3 , 900kg / m 3 .

[0121] S6. The report generation module integrates the three - dimensional coordinates of the leakage point, the LI value, the influence radius R, and the repair suggestions into a structured report and uploads it to the monitoring center.

[0122] Furthermore, in the above - mentioned technical solution, the repair suggestions include:

[0123] When LI > LI1, stop the transmission urgently;

[0124] When LI2 < LI ≤ LI1, strengthen the monitoring;

[0125] When LI3 < LI ≤ LI2, prompt potential risks.

[0126] It is important to understand that LI1 is the threshold for major leakage risk, LI2 is the threshold for moderate leakage risk, and LI3 is the threshold for potential leakage risk. The values ​​of LI1, LI2, and LI3 are set as follows: the initial values ​​are dynamically generated based on historical leakage events and operating condition data; LI1 is taken from the top 5% to 10% percentile of LI values ​​in historical leakage events to ensure that only extreme leakage events trigger a shutdown; LI2 is taken from the median of historical data to distinguish between moderate and high risk; LI3 is based on the upper limit of LI fluctuation under normal operating conditions (e.g., mean + 3 standard deviation) to capture minor anomalies; to adapt to complex environments, LI data of new leakage events are injected periodically, such as monthly, to recalculate percentiles to iterate LI1 and LI2, while LI3 is adjusted downwards in conjunction with real-time weather conditions, such as strong winds and high temperatures, to improve sensitivity.

[0127] The specific procedure for emergency shutdown is as follows: The monitoring center sends an encrypted shutdown command to the pipeline SCADA system, driving the upstream valve to close and cut off the flow within ≤3 minutes; At the same time, the on-duty engineer retrieves the three-dimensional coordinates of the leak point and the real-time video stream from the drone for double confirmation. If the verification confirms a real leak within 10 minutes, the fire and environmental protection departments will be coordinated to activate the emergency plan; otherwise, the false alarm will be deactivated.

[0128] The specific process for enhanced monitoring is as follows: the frequency of drone inspections is increased from the usual 4 hours / time to 30 minutes / time, the flight altitude is reduced to 10-20 meters to improve data resolution, the infrared sampling rate is increased to 20Hz and the acoustic array is expanded to the infrasound band.

[0129] This invention provides, for example Figure 2 The pipeline leak monitoring system based on drone inspection shown includes the following modules:

[0130] The pipeline status data acquisition module is used to inspect pipelines along a preset route using a drone equipped with multiple sensors, and to collect multimodal raw data within a certain flight altitude range to form a pipeline status dataset; the pipeline status dataset includes infrared thermal imaging data, air pressure fluctuation data, and acoustic signature data.

[0131] The pipeline status data preprocessing module is used to perform spatiotemporal alignment, noise filtering, and data standardization on the pipeline status dataset.

[0132] The pipeline status data analysis module is used to analyze the preprocessed pipeline status dataset, calculate the temperature anomaly coefficient K1, air pressure fluctuation coefficient K2 and acoustic fingerprint matching coefficient K3, input the temperature anomaly coefficient K1, air pressure fluctuation coefficient K2 and acoustic fingerprint matching coefficient K3 into the pipeline leakage assessment model, and output the pipeline leakage index LI.

[0133] The leakage detection module is used to compare the pipeline leakage index LI value with a preset threshold θ. When LI>θ continuously within a certain sampling period, a leakage warning is triggered and the location module is activated.

[0134] The leak location module is used to calculate the three-dimensional coordinates (x, y, z) and influence radius R of the leak point based on the pipeline leak index LI value, combined with the UAV's GPS coordinates, the sound wave arrival time difference Δt at the moment of sudden air pressure change, and the wind speed vector W. a ;

[0135] The report generation module is used to generate the three-dimensional coordinates of the leak point, the LI value, and the radius of influence R. a The repair suggestions are integrated into a structured report and uploaded to the monitoring center.

[0136] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for monitoring pipeline leaks based on unmanned aerial vehicle (UAV) inspection, characterized in that, It includes the following steps: S1. The pipeline status data acquisition module uses a drone equipped with multiple sensors to inspect the pipeline along a preset route, and acquires multi-modal raw data within a certain flight altitude range to form a pipeline status data set; the pipeline status data set includes infrared thermal imaging data, air pressure fluctuation data, and voiceprint feature data; S2. The pipeline status data preprocessing module performs spatio-temporal alignment, noise filtering, and data standardization processing on the pipeline status data set; S3. The pipeline status data analysis module analyzes the preprocessed pipeline status data set, calculates the temperature anomaly coefficient K1, the air pressure fluctuation coefficient K2, and the voiceprint matching coefficient K3, inputs the temperature anomaly coefficient K1, the air pressure fluctuation coefficient K2, and the voiceprint matching coefficient K3 into the pipeline leakage assessment model, and outputs the pipeline leakage index LI; S4. The leakage determination module compares the pipeline leakage index LI value with a preset threshold θ. When LI>θ continuously within a certain sampling period, a leakage warning is triggered, and the positioning module is activated; S5. The leak location module calculates the three-dimensional coordinates (x, y, z) and influence radius R of the leak point based on the pipeline leakage index LI value, combined with the UAV's GPS coordinates, the sound wave arrival time difference Δt at the moment of sudden air pressure change, and the wind speed vector W. a ; S6. The report generation module will generate the three-dimensional coordinates of the leak point, the LI value, and the radius of influence R. a The repair suggestions are integrated into a structured report and uploaded to the monitoring center.

2. The pipeline leakage monitoring based on UAV inspection according to claim 1, characterized in that, The infrared thermal imaging data includes the maximum temperature difference between adjacent frames, the ambient reference temperature, and the proportion of the effective area of the abnormally high temperature area in a single-frame infrared image; the air pressure fluctuation data includes the instantaneous air pressure value and the air pressure gradient change rate; the voiceprint feature data includes the spectral amplitude.

3. The method for monitoring pipeline leaks based on unmanned aerial vehicle (UAV) inspection according to claim 1, characterized in that, The temperature anomaly coefficient K1 is specifically: Where, ΔT max The maximum temperature difference between adjacent frames, T0 is the ambient reference temperature, and T crit Where A is the critical temperature for thermal deformation of the pipe material, A0 is the critical threshold for determining the proportion of the high-temperature region in the pipe, m is the area weight attenuation coefficient, and e is the natural constant. radio This represents the effective area percentage of abnormally high-temperature regions in a single frame of infrared image.

4. The method for monitoring pipeline leaks based on unmanned aerial vehicle (UAV) inspection according to claim 1, characterized in that, The air pressure fluctuation coefficient K2 is specifically: Where P(t) is the instantaneous air pressure value at second t, σ P(t) Let σt be the standard deviation of air pressure in second t, σ0 be the threshold standard deviation under normal operating conditions, and η be the standard deviation of air pressure in second t. P(t) η is the rate of change of the pressure gradient. max is the historical maximum gradient value, and tanh is the hyperbolic tangent function.

5. The method for monitoring pipeline leaks based on unmanned aerial vehicle (UAV) inspection according to claim 1, characterized in that, The voiceprint matching coefficient K3 is specifically: Where f is the discrete frequency point of spectral analysis, S(f) is the spectral amplitude of the real-time collected acoustic signal at frequency f, R(f) is the spectral amplitude of the preset pipeline leakage voiceprint feature template at frequency f, f1 is the lower limit frequency of the characteristic frequency band, f2 is the upper limit frequency of the characteristic frequency band, L is the propagation distance, a is the distance attenuation constant, and e is the natural constant.

6. The method for monitoring pipeline leaks based on UAV inspection according to claim 1, characterized in that, The pipeline leakage assessment model is specifically: Where LI is the pipeline leakage index, K1 is the temperature anomaly coefficient, K2 is the air pressure fluctuation coefficient, K3 is the acoustic signature matching coefficient, α is the pressure attenuation factor, and P... thr Where β is the pressure threshold, λ is the acoustic attenuation factor, and A is the global risk factor. thr η is the acoustic threshold, η is the overall sensitivity, and e is the natural constant.

7. The method for monitoring pipeline leaks based on unmanned aerial vehicle (UAV) inspection according to claim 1, characterized in that, The calculation method of the three-dimensional coordinates of the leakage point is: Where x0 is the horizontal coordinate of the UAV's GPS, y0 is the horizontal coordinate of the UAV's GPS, and v x Let v be the linear velocity component of the object in the x-direction. y W represents the linear velocity component of the object in the y-direction. x W represents the angular velocity component of the object about the x-axis. y Let H be the angular velocity component of the object about the y-axis, H be the flight altitude, ρ be the air density, g be the acceleration due to gravity, and P be the acceleration due to gravity. d P0 is the ambient atmospheric pressure, and Δt is the time difference of sound wave arrival at the moment of sudden change in air pressure.

8. The method for monitoring pipeline leaks based on unmanned aerial vehicle (UAV) inspection according to claim 1, characterized in that, The radius of influence R a The calculation formula is: The radius of influence R a The calculation formula is: Where R0 is the reference influence radius, k R ω is the leakage exponential amplification factor, ||W|| is the wind speed correction factor, and ||W|| is the wind speed modulus. gas Let be the initial velocity of the gas injection, and e be the natural constant.

9. The method for monitoring pipeline leaks based on unmanned aerial vehicle (UAV) inspection according to claim 1, characterized in that, The repair suggestions include: When LI>LI1, stop the transmission urgently; When LI2<LI≤LI1, strengthen the monitoring; When LI3<LI≤LI2, prompt potential risks.

10. A pipeline leakage monitoring system based on drone inspection specifically includes the following modules: [[ID= ​ ​ The leakage detection module is used to compare the pipeline leakage index LI value with a preset threshold θ. When LI>θ continuously within a certain sampling period, a leakage warning is triggered and the location module is activated. The leak location module is used to calculate the three-dimensional coordinates (x, y, z) and influence radius R of the leak point based on the pipeline leak index LI value, combined with the UAV's GPS coordinates, the sound wave arrival time difference Δt at the moment of sudden air pressure change, and the wind speed vector W. a ; The report generation module is used to generate the three-dimensional coordinates of the leak point, the LI value, and the radius of influence R. a The repair suggestions are integrated into a structured report and uploaded to the monitoring center.

Images