CN-121999801-A - Heating pipeline leakage online diagnosis and positioning system and method based on acoustic sensing array and AI algorithm

CN121999801ACN 121999801 ACN121999801 ACN 121999801ACN-121999801-A

Abstract

The invention relates to a heating pipeline leakage on-line diagnosis and positioning system and method based on an acoustic sensing array and an AI algorithm, belonging to the field of pipeline safety detection; the system comprises a distributed sensing network, a cloud platform and a data fusion algorithm, wherein the distributed sensing network comprises a plurality of acoustic sensing devices distributed at intervals along a heat supply pipeline, each acoustic sensing device is used for collecting sound wave signals, an edge computing unit is in communication connection with the acoustic sensing devices and used for preprocessing the sound wave signals to obtain feature vectors, the cloud platform is in communication connection with the edge computing units and is internally provided with a neural network model in a preset mode, and the cloud platform is configured to comprehensively process information of the acoustic sensing devices through the data fusion algorithm so as to diagnose and position leakage points on the heat supply pipeline.

Inventors

Song Guowai
GAO WENJING
LI XIAOTAO
REN YAN
YANG DONGJIANG
HAN GUANHENG
LIU HONGBIN
JIANG KAI
WANG YUHANG
Bai Shanshui
GUO YULONG
WANG ZHUO

Assignees

中国能源建设集团山西省电力勘测设计院有限公司

Dates

Publication Date: 20260508
Application Date: 20260113

Claims (10)

1. An on-line diagnosis and positioning system for heat supply pipeline leakage based on an acoustic sensing array and an AI algorithm is characterized by comprising the following steps: The distributed sensing network comprises a plurality of acoustic sensing devices (1) which are distributed at intervals along a heat supply pipeline (2), wherein each acoustic sensing device (1) is used for acquiring sound wave signals; An edge computing unit (3) in communication with the acoustic sensing device (1) for preprocessing acoustic signals to obtain feature vectors; the cloud platform (4) is in communication connection with the edge computing unit (3), and a neural network model is preset in the cloud platform; the cloud platform (4) is configured to comprehensively process information of the plurality of acoustic sensing devices (1) through a data fusion algorithm so as to realize diagnosis and positioning of leakage points on the heat supply pipeline (2), and specifically comprises the following steps: The cloud platform (4) is used for fusing the feature vectors from the plurality of edge computing units (3), and judging whether leakage points exist in the heat supply pipeline (2) through the neural network model; when the heat supply pipeline (2) has a leakage point, the cloud platform (4) is further used for calculating the position of the leakage point through a data fusion positioning algorithm based on the time difference of arrival of sound wave signals acquired by the at least two acoustic sensing devices (1).
2. The heating pipeline leakage on-line diagnosis and positioning system based on the acoustic sensor array and the AI algorithm as set forth in claim 1, wherein the acoustic sensor array is a triangular array composed of three sensors (5).
3. The heating pipeline leakage online diagnosis and positioning system based on the acoustic sensing array and the AI algorithm of claim 1, wherein the preprocessing comprises noise reduction processing and feature extraction, the noise reduction processing comprises power frequency interference suppression, feature frequency band extraction and adaptive denoising to remarkably improve the signal-to-noise ratio of the sound wave signals, and the feature extraction comprises extracting multidimensional features of the sound wave signals from a time domain, a frequency domain and a time-frequency domain to form feature vectors.
4. The heating pipeline leakage online diagnosis and positioning system based on the acoustic sensing array and the AI algorithm of claim 1, wherein the feature vector is 96-dimensional, and comprises 16-dimensional time domain statistical features, 32-dimensional frequency domain features and 48-dimensional time-frequency features.
5. The heating pipeline leakage online diagnosis and positioning system based on the acoustic sensing array and the AI algorithm as set forth in claim 1, wherein the neural network model is an improved convolutional neural network comprising an input layer, a feature extraction layer, a feature fusion layer and an output layer, wherein the improved convolutional neural network is a YOLOv network for introducing an attention mechanism, and an activation function is a Mish function.
6. The heating pipeline leakage online diagnosis and positioning system based on the acoustic sensing array and the AI algorithm according to claim 1, wherein the data fusion positioning algorithm is an algorithm for calculating the time difference of arrival of an acoustic wave signal to an acoustic sensing device (1) based on a generalized cross correlation algorithm and calculating the position of a leakage point through a hyperbolic positioning model by combining the propagation speed of the acoustic wave and the distance between two acoustic sensor devices which acquire the acoustic wave signal of the same leakage point.
7. The heating pipeline leakage on-line diagnosis and positioning system based on the acoustic sensing array and the AI algorithm of claim 6, wherein the generalized cross-correlation algorithm is a phase transformation weighting process, and the frequency range is 1kHz-8kHz.
8. The heating pipeline leakage on-line diagnosis and positioning system based on the acoustic sensing array and the AI algorithm as claimed in claim 1, wherein the distance between two adjacent acoustic sensing devices (1) is 50-100 meters.
9. An on-line diagnosis and positioning method for heat supply pipeline leakage based on an acoustic sensor array and an AI algorithm, which adopts the on-line diagnosis and positioning system for heat supply pipeline leakage based on the acoustic sensor array and the AI algorithm as set forth in any one of claims 1-8, and is characterized by comprising the following steps: s1, acquiring sound wave signals corresponding to a heat supply pipeline (2) through a plurality of acoustic sensor devices which are fixedly arranged at intervals along the heat supply pipeline (2); s2, extracting multidimensional features from the time domain, the frequency domain and the time-frequency domain of the acoustic wave signal after noise reduction to form feature vectors; step S3, the neural network model judges whether a leakage point exists in the heat supply pipeline (2) based on the received characteristic vector obtained through the edge calculation unit (3); and S4, when the neural network model identifies the current acoustic signal and judges that the heat supply pipeline (2) has the leakage point, the neural network model calculates the position of the heat supply pipeline (2) where the leakage point is positioned based on the time difference of arrival of the acoustic signals transmitted by the at least two acoustic sensing devices (1), and a leakage point positioning result is obtained.
10. The on-line diagnosis and positioning method for heat supply pipeline leakage based on an acoustic sensor array and an AI algorithm of claim 9, wherein when the neural network model identifies the current acoustic wave signal to determine that the heat supply pipeline (2) has the leakage point, the method for obtaining the positioning result of the leakage point comprises the following steps: Step S41, the neural network model transmits a signal of a leakage point of the heat supply pipeline (2) to the acoustic sensor device and outputs an early warning signal, and simultaneously, the time stamp of the leakage point signal reaching the acoustic sensor device is recorded and uploaded to the cloud platform (4) through the communication module; Step S42, after a neural network model preset in the cloud platform (4) is based on acoustic signals reported by at least two acoustic sensing devices (1) aiming at the same leakage event, calculating the time difference that the acoustic signals sent by the leakage points reach the acoustic sensing devices (1) which acquire the same leakage event based on a generalized cross-correlation algorithm; S43, correcting the sound wave propagation speed by combining a neural network model preset in the cloud platform (4) with the temperature and pressure parameters of the heating pipeline (2) where the leakage point is located; s44, inputting the time difference and the corrected sound wave propagation speed into a hyperbola positioning model, and calculating the space coordinates of the leakage point; and step S45, transmitting the space coordinates of the leakage points and the early warning signals to a monitoring center for graphical display.

Description

Heating pipeline leakage online diagnosis and positioning system and method based on acoustic sensing array and AI algorithm Technical Field The invention provides a heating pipeline leakage online diagnosis and positioning system and method based on an acoustic sensor array and an AI algorithm, and belongs to the technical field of pipeline safety detection. Background The urban central heating system is an important component of urban infrastructure, and the scale of a heating pipeline network is continuously enlarged along with the acceleration of the urban process. However, in the long-term operation of the pipeline, leakage frequently occurs due to factors such as aging, corrosion, external disturbance and the like, and the pipeline becomes a safe operation bottleneck. The leakage not only wastes heat energy and water resources to cause economic loss, but also can cause secondary accidents such as pavement collapse and the like, thereby threatening the environment and personnel safety. The existing heat supply pipeline leakage detection technology mainly comprises manual inspection, pressure detection, optical fiber sensing, acoustic detection, infrared thermal imaging and the like, but each has a short plate: the manual inspection relies on experience, the efficiency is low, and the omission factor is high; The pressure detection positioning precision is poor, and the sensitivity to micro leakage is insufficient; The optical fiber sensing has high cost, is easy to be interfered, and has weak environment noise resistance in acoustic detection; Infrared thermal imaging is greatly affected by weather and other environments, and has poor applicability to buried pipelines. In conclusion, the prior art generally has the problems of insufficient sensitivity, poor environmental adaptability, limited positioning precision, unbalanced cost and performance and the like. Therefore, research and development of a detection system with high sensitivity, strong anti-interference, high-precision positioning and reasonable cost is a difficult problem to be solved urgently, and has important practical significance and application value. Disclosure of Invention The invention provides a heating pipeline leakage on-line diagnosis and positioning system and method based on an acoustic sensing array and an AI algorithm, aiming at solving the technical problems of insufficient detection sensitivity, weak anti-interference capability and low positioning accuracy of the existing pipeline leakage detection system. In order to solve the technical problems, the invention adopts the technical scheme that the on-line diagnosis and positioning system for the leakage of the heating pipeline based on the acoustic sensor array and the AI algorithm is characterized in that a plurality of acoustic sensors are adopted, and acoustic signals acquired by a plurality of acoustic sensors distributed along the heating pipeline are comprehensively processed through a data fusion algorithm so as to improve the reliability of the diagnosis and the positioning accuracy of the leakage points. The system comprises: the distributed sensing network comprises a plurality of acoustic sensing devices which are distributed at intervals along a heat supply pipeline, wherein each acoustic sensing device is used for acquiring sound wave signals; the edge computing unit is in communication connection with the acoustic sensing device and is used for preprocessing the acoustic wave signals to obtain feature vectors; the cloud platform is in communication connection with the edge computing unit, and a neural network model is preset in the cloud platform; The cloud platform is configured to comprehensively process information of a plurality of acoustic sensing devices through a data fusion algorithm so as to realize diagnosis and positioning of leakage points on a heat supply pipeline, and specifically comprises the following steps: the cloud platform is used for fusing the feature vectors from the plurality of edge computing units and judging whether leakage points exist in the heat supply pipeline or not through the neural network model; when a leakage point exists in the heat supply pipeline, the cloud platform is further used for calculating the position of the leakage point through a data fusion positioning algorithm based on the time difference of arrival of sound wave signals acquired by at least two acoustic sensing devices. Further, the acoustic sensor array is a triangular array of three sensors. Further, the preprocessing comprises noise reduction processing and feature extraction, the noise reduction processing comprises power frequency interference suppression, feature frequency band extraction and self-adaptive denoising so as to remarkably improve the signal to noise ratio of the sound wave signal, and the feature extraction comprises the step of extracting multidimensional features of the sound wave signal from a time domain, a frequency domain and