CN-121993747-A - Distributed acoustic sensing and digital twin-based intelligent diagnosis system and method for leakage of heat supply pipeline

CN121993747ACN 121993747 ACN121993747 ACN 121993747ACN-121993747-A

Abstract

The invention discloses a distributed acoustic sensing and digital twin-based intelligent diagnosis system and method for heat supply pipeline leakage, belongs to the field of heat supply pipeline leakage detection, and solves the problems of low precision, low pertinence, high false alarm rate and lack of self-learning and evolution capability of an existing heat supply pipeline detection method. The system also has the open-set self-learning capability, can automatically identify and learn an unknown fault mode, and is applied to the leakage detection of the heat pipe.

Inventors

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

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 (9)

1. An intelligent heat supply pipeline leakage diagnosis system based on distributed acoustic sensing and digital twinning is characterized by comprising the following components: the high-synchronization distributed acoustic sensing network comprises a plurality of acoustic sensing devices distributed along a heat supply pipeline; the edge intelligent diagnosis unit comprises a cooperative signal enhancement module, wherein the cooperative signal enhancement module is used for receiving synchronous original sound wave signals from a plurality of acoustic sensing devices and generating enhanced sound wave signals pointing to the potential leakage direction; The multi-dimensional characteristic extraction module is used for extracting high-dimensional characteristic vectors in the enhanced acoustic wave signals, and the multi-dimensional intelligent recognition module is used for outputting binary classification of leakage existence, multi-classification of leakage types and regression estimation of leakage quantity according to the high-dimensional characteristic vectors; The cloud platform and the digital twin body unit comprise a high-precision positioning module and an open-collection self-learning module, wherein the high-precision positioning module is used for positioning the position of the leakage point based on the enhanced acoustic wave signal; The high-precision positioning module is in communication connection with a virtual checking and confidence evaluating module, and the virtual checking and confidence evaluating module is used for constructing an acoustic digital twin model corresponding to the heat supply pipeline so as to simulate the position and type of a leakage point to generate a simulated acoustic signal and output a diagnosis confidence level; the open-set self-learning module is used for continuously monitoring and clustering unrecognized abnormal acoustic modes, and the new labeling data is utilized to trigger incremental update of the multi-task intelligent recognition module.
2. The intelligent diagnosis system for heat supply pipeline leakage based on distributed acoustic sensing and digital twinning of claim 1 is characterized in that the acoustic sensing device comprises a data acquisition module and an acoustic array module formed by a plurality of MEMS microphones, the data acquisition module is used for acquiring acoustic wave signals transmitted by the acoustic array module, the data acquisition module is sequentially and electrically connected with a high-precision clock synchronization module, a data processing module and a communication module, microsecond time synchronization is achieved among the acoustic sensing devices through a preset precision clock synchronization protocol, and timing consistency of data acquired by the multiple acoustic sensing devices is ensured.
3. The intelligent diagnosis system for heat supply pipeline leakage based on distributed acoustic sensing and digital twinning as set forth in claim 1, wherein the multi-task deep learning model adopts a framework based on meta-learning, and a feature extraction layer of the multi-task deep learning model can quickly adapt to a new leakage mode through a small number of new samples on the basis of pre-training, thereby realizing quick transfer learning of few samples and even zero samples.
4. The distributed acoustic sensing and digital twinning based intelligent heating conduit leakage diagnostic system according to claim 1 wherein the high-dimensional feature vector comprises conventional time-frequency features, and higher-order spectral features and acoustic emission modal features for characterizing leakage physical characteristics.
5. The distributed acoustic sensing and digital twinning based intelligent heating pipeline leakage diagnostic system according to claim 1, wherein the simulated acoustic signals comprise sound pressure level, spectral characteristics and acoustic wave modes.
6. The distributed acoustic sensing and digital twinning based intelligent thermal pipeline leakage diagnostic system according to claim 1, wherein the multiple classifications of leakage types include at least one or more of sand holes, cracks, and corrosive perforations.
7. The intelligent diagnosis system for heat supply pipeline leakage based on distributed acoustic sensing and digital twinning according to claim 1, further comprising a dynamic credibility fusion module, wherein the dynamic credibility fusion module is used for receiving the output result from the multi-task intelligent identification module, the diagnosis confidence level output by the virtual verification and confidence assessment module and historical maintenance data, and outputting the priority of comprehensive credibility score and operation and maintenance suggestion through a D-S evidence theory or Bayesian inference model.
8. A distributed acoustic sensing and digital twin based intelligent diagnosis method for heat supply pipeline leakage, which adopts the distributed acoustic sensing and digital twin based intelligent diagnosis system for heat supply pipeline leakage as claimed in any one of claims 1-7, and is characterized by comprising the following steps: step S1, acquiring sound wave signals of a heat supply pipeline based on a high-synchronization distributed acoustic sensor network; S2, cooperatively enhancing the acquired acoustic wave signals based on the edge intelligent diagnosis unit, and extracting high-dimensional feature vectors in the enhanced acoustic wave signals; Step S3, performing leakage detection of a heat supply pipeline, classification of leakage points and evaluation operation of the leakage points based on the high-dimensional feature vector; S4, the high-precision positioning module positions the positions of the leakage points based on the enhanced acoustic signals, and a virtual verification and confidence evaluation module is started to perform virtual verification; And S5, generating a comprehensive diagnosis report and a reliability score based on the virtual verification result and the recognition result of the multi-task intelligent recognition module.
9. The intelligent diagnosis method of heat supply pipeline leakage based on distributed acoustic sensing and digital twinning according to claim 8, wherein in step S5, when the multitasking intelligent recognition module recognizes and records the unknown abnormal signal for a preset number of times, the unknown abnormal signal is transmitted to the open-collection self-learning module, the open-collection self-learning module clusters and marks the continuously occurring unknown abnormal signal, and the open-collection self-learning module self-learning process is triggered.

Description

Distributed acoustic sensing and digital twin-based intelligent diagnosis system and method for leakage of heat supply pipeline Technical Field The invention relates to the technical field of heat supply pipeline leakage detection, in particular to an intelligent heat supply pipeline leakage diagnosis system and method based on distributed acoustic sensing and digital twinning. Background Urban district heating systems are an important component of urban lifeline infrastructure. Along with the expansion of the scale of the pipe network and the growth of the service life, the risk of leakage accidents caused by aging and corrosion is obviously increased, so that huge energy waste and economic loss are caused, and serious safety and environmental accidents are more likely to be caused. However, the existing heat supply pipeline leakage monitoring technology also has the following bottlenecks: 1. The sensing capability is insufficient, the traditional acoustic or negative pressure wave method is easy to be interfered by complex environmental noise, the detection sensitivity to tiny leakage (such as sand holes) is low, and the signal to noise ratio is poor. 2. The diagnosis function is single, most systems can only realize binary judgment of 'whether leakage exists', and the leakage type and the evaluation severity degree cannot be identified, so that the operation and maintenance decision is lack of pertinence. 3. The reliability is doubtful, the false alarm rate is high, a large number of invalid alarms consume human resources, and the reliability of the system is reduced. 4. The intelligent degree is low, the system model is fixed, the system model cannot adapt to pipeline state change and novel fault modes, the self-learning and evolution capability is lacked, and the maintenance cost of the whole life cycle is high. Therefore, developing an intelligent leakage monitoring system capable of realizing high sensitivity sensing, fine diagnosis, high reliability early warning and self-evolution capability has become an urgent need in the industry. Disclosure of Invention The invention provides an intelligent diagnosis system and method for heat supply pipeline leakage based on distributed acoustic sensing and digital twin, which can not only locate leakage points with high precision, but also realize leakage type identification, severity assessment and virtual check false alarm, and has continuous self-learning capability, aiming at solving the technical problems of low precision, low pertinence, high false alarm rate and lack of self-learning and evolution capability of the existing heat supply pipeline detection method. In order to solve the technical problems, the invention adopts the technical scheme that the intelligent heat supply pipeline leakage diagnosis system based on distributed acoustic sensing and digital twinning comprises: the high-synchronization distributed acoustic sensing network comprises a plurality of acoustic sensing devices distributed along a heat supply pipeline; the edge intelligent diagnosis unit comprises a cooperative signal enhancement module, wherein the cooperative signal enhancement module is used for receiving synchronous original sound wave signals from a plurality of acoustic sensing devices and generating enhanced sound wave signals pointing to the potential leakage direction; The multi-dimensional characteristic extraction module is used for extracting high-dimensional characteristic vectors in the enhanced acoustic wave signals, and the multi-dimensional intelligent recognition module is used for outputting binary classification of leakage existence, multi-classification of leakage types and regression estimation of leakage quantity according to the high-dimensional characteristic vectors; The cloud platform and the digital twin body unit comprise a high-precision positioning module and an open-collection self-learning module, wherein the high-precision positioning module is used for positioning the position of the leakage point based on the enhanced acoustic wave signal; The high-precision positioning module is in communication connection with a virtual checking and confidence evaluating module, and the virtual checking and confidence evaluating module is used for constructing an acoustic digital twin model corresponding to the heat supply pipeline so as to simulate the position and type of a leakage point to generate a simulated acoustic signal and output a diagnosis confidence level; the open-set self-learning module is used for continuously monitoring and clustering unrecognized abnormal acoustic modes, and the new labeling data is utilized to trigger incremental update of the multi-task intelligent recognition module. The acoustic sensing device comprises a data acquisition module and an acoustic array module formed by a plurality of MEMS microphones, wherein the data acquisition module is used for acquiring acoustic signals transmitted by the acoustic array module,