CN-122017039-A - Acoustic emission nonmetallic pipeline damage detection method based on intelligent sensor

CN122017039ACN 122017039 ACN122017039 ACN 122017039ACN-122017039-A

Abstract

The invention discloses an acoustic emission nonmetallic pipeline damage detection method based on an intelligent sensor, and particularly relates to the field of nondestructive detection; the method comprises the steps of arranging intelligent acoustic emission sensor nodes with built-in multi-modules on the outer wall of a non-metal pipeline, constructing a distributed self-organizing network monitoring system, achieving microsecond time synchronization, extracting effective waveforms through pre-amplification, self-adaptive filtering and background noise suppression after acoustic emission signals are collected by all the nodes, extracting time domain-frequency domain characteristic parameters through an edge computing unit, completing preliminary identification of damage types by utilizing a lightweight convolutional neural network, locating damage sources by the edge gateway based on arrival time difference corrected by frequency dispersion characteristics in multi-node cooperative response, eliminating abnormal points by combining signal attenuation trend and confidence, finally fusing detection data by a cloud platform, carrying out damage evolution trend analysis and residual life prediction, and triggering multi-stage early warning.

Inventors

WEI JUNXIAO
WEI WEI
CUI YUNLONG
LI XIANGZHENG
ZHANG YUHU
WANG YUHAN

Assignees

山东泰阳特种设备检测科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260403

Claims (10)

1. The acoustic emission nonmetallic pipeline damage detection method based on the intelligent sensor is characterized by comprising the following steps of: A1, arranging a plurality of intelligent acoustic emission sensor nodes on the outer wall of a nonmetallic pipeline at equal intervals along the axial direction, wherein a piezoelectric film sensor, a temperature compensation module, an edge calculation unit and a wireless communication module are arranged in each node, microsecond synchronization is realized among the nodes through a time synchronization protocol, and a distributed self-networking monitoring system is formed; A2, each intelligent sensor node acquires an acoustic emission signal in real time, and after the acoustic emission signal is amplified by a preamplifier, the acoustic emission signal is subjected to self-adaptive filtering and background noise suppression by an edge computing unit, and an effective acoustic emission waveform is extracted; a3, extracting time domain characteristic parameters and frequency domain characteristic parameters of the acoustic emission signals by an edge computing unit, constructing a characteristic classification model based on a lightweight convolutional neural network, and realizing preliminary identification of damage types at a node end; a4, when a plurality of adjacent nodes detect effective acoustic emission events at the same time, each node uploads an event time stamp, a feature vector and a preliminary classification result to an edge gateway, and the edge gateway corrects a wave velocity model based on an arrival time difference method by combining the frequency dispersion characteristics of acoustic emission signals in a nonmetallic pipeline to calculate the position of a damage source; and A5, uploading the damage type, the position information and the intensity level to a cloud management platform by the edge gateway, carrying out damage evolution trend analysis and residual life prediction by combining historical data and pipeline operation parameters, and triggering multi-stage early warning according to a preset threshold value.
2. The method for detecting damage to an acoustic emission nonmetallic pipeline based on an intelligent sensor as set forth in claim 1, wherein in the intelligent acoustic emission sensor node, the temperature compensation module is composed of a high-precision digital temperature sensor and is used for collecting the surface temperature of the pipeline in real time and correcting nonlinear influences of the temperature on the propagation speed of acoustic emission signals and the sensitivity of the sensor.
3. The method for detecting the damage of the acoustic emission nonmetallic pipeline based on the intelligent sensor according to claim 1, wherein in the A2, an improved wavelet threshold denoising algorithm is adopted for self-adaptive filtering, a threshold function is dynamically adjusted by combining typical noise spectrum characteristics of the pipeline, and in the improved wavelet threshold denoising algorithm, an improved soft threshold function is adopted for threshold processing.
4. The method for detecting the damage of the acoustic emission nonmetallic pipeline based on the intelligent sensor according to claim 1, wherein in the A3, a light convolution neural network adopts a depth separable convolution structure, real-time reasoning is carried out on an embedded processor at a node end, the network structure comprises three depth separable convolution blocks, each convolution block comprises a depth convolution layer and a point-by-point convolution layer, a normalization layer and a ReLU activation function are connected, and damage type probability distribution is output through a global average pooling layer and a Softmax classifier.
5. The method for detecting the damage of the acoustic emission nonmetallic pipeline based on the intelligent sensor as claimed in claim 4, wherein the model training stage of the lightweight convolutional neural network adopts a migration learning strategy, firstly pretrains the network by using a labeling sample collected in a laboratory environment, and then fine-adjusts the labeling sample under the actual working condition on site to adapt the model to the specific pipeline material and the environment characteristic.
6. The method for detecting the damage of the acoustic emission nonmetallic pipeline based on the intelligent sensor according to claim 1, wherein in A4, a frequency dispersion characteristic correction wave speed model is established in advance through laboratory tests, mapping relations between different frequency components of acoustic emission signals and propagation speeds are established, in arrival time difference calculation, weighted correction is conducted on centroid frequencies of signals received by all nodes, and a distance difference equation is established by adopting weighted average correction speeds.
7. The method for detecting acoustic emission nonmetallic pipeline damage based on the intelligent sensor, which is characterized by removing abnormal locating points based on signal attenuation trend and node response confidence, comprises calculating theoretical amplitude according to an exponential attenuation relation between signal amplitude and propagation distance, comparing the theoretical amplitude with actual detection amplitude to obtain relative errors, removing nodes with errors exceeding a preset threshold, and meanwhile, marking low-confidence nodes based on consistency of node end classification results and classification confidence, and constructing a geometric constraint model to remove the abnormal locating points.
8. The method for detecting the damage of the acoustic emission nonmetallic pipeline based on the intelligent sensor as claimed in claim 6, wherein the method for eliminating abnormal points by combining the confidence coefficient specifically comprises a secondary rechecking mechanism: When the damage source position confidence calculated by the edge gateway preliminarily is lower than a preset threshold, the sensor nodes in the target area are scheduled to acquire data again at a sampling rate higher than an initial set value; The edge gateway calculates a positioning result for a plurality of times by utilizing a sliding window method, and extracts the statistical mean and variance of the positioning coordinates for a plurality of times; and taking the mean coordinate with the variance smaller than or equal to the limit value as the final output damage source position.
9. The method for detecting damage to an acoustic emission nonmetallic pipeline based on an intelligent sensor as set forth in claim 6, wherein when a positioning result is marked as a low confidence event, the edge gateway triggers a secondary rechecking mechanism of adjacent nodes, instructs 3 nodes near a damage source to re-acquire signals at a higher sampling rate, calculates a statistical mean and variance through multiple rounds of positioning, and outputs a damage position after confidence weighted fusion.
10. The method for detecting the damage of the acoustic emission nonmetallic pipeline based on the intelligent sensor according to claim 1, wherein in the A5, the damage evolution trend analysis adopts a sliding window autoregressive model to fit a damage intensity time sequence, and a cumulative damage index is introduced to judge a damage evolution stage; The method comprises the steps of predicting the residual life, calculating the residual cycle times and the residual service time of a polyethylene pipeline by adopting a crack propagation model based on a Paris formula, estimating the residual life by adopting a cumulative damage mechanical model and through residual rigidity for a fiber reinforced composite pipe, and comprehensively judging each stage of early warning according to a damage strength grade, a cumulative damage index, a residual life prediction result and event positioning confidence level by adopting four stages of blue early warning, yellow early warning, orange early warning and red early warning, wherein different triggering conditions and treatment measures are respectively corresponding to each other.

Description

Acoustic emission nonmetallic pipeline damage detection method based on intelligent sensor Technical Field The invention relates to the technical field of nondestructive testing, in particular to an acoustic emission nonmetallic pipeline damage detection method based on an intelligent sensor. Background The nonmetallic pipeline has the advantages of corrosion resistance, light weight, low cost, long service life and the like, and is widely applied to the fields of municipal water supply, gas transportation, chemical medium transportation, ocean engineering and the like. However, non-metallic materials are susceptible to failure modes such as creep deformation, impact damage, crack growth, aging degradation and the like in the service process, and once leakage or cracking accidents occur, serious economic loss and environmental hazard are caused. Therefore, the method has important engineering significance in real-time and online damage detection and state evaluation of the nonmetallic pipeline. The acoustic emission detection technology has become an important technical means for the health monitoring of nonmetallic pipelines because of the characteristics of high sensitivity, passive monitoring, strong real-time performance and the like of the acoustic emission detection technology on the dynamic damage inside the material. However, the conventional nonmetallic pipeline detection technology based on acoustic emission still has the following problems in practical application: Firstly, nonmetallic pipeline materials (such as polyethylene, glass fiber reinforced plastic and the like) have serious attenuation on high-frequency components of acoustic emission signals, and the traditional piezoelectric sensor has low signal-to-noise ratio and is easily influenced by complex background noise such as on-site pump valve vibration, fluid turbulence, electromagnetic interference and the like, so that the effective event detection omission ratio and false alarm ratio are high. Secondly, the existing acoustic emission detection system mostly adopts a centralized data acquisition and processing architecture, the sensor is connected to a central acquisition host through a long cable, the system wiring is complex, the expansibility is poor, a large amount of original data is required to be transmitted to a central processor, the data transmission quantity is large, the real-time performance is limited, and long-distance and large-range distributed monitoring is difficult to realize. Thirdly, the damage positioning depends on the propagation speed of acoustic emission signals in a pipeline medium, the wave speed is regarded as a constant by the traditional method, but the nonmetallic material presents obvious dispersion characteristics, namely signals with different frequency components have different propagation speeds, so that the arrival time difference positioning method based on fixed wave speed has larger error, and the positioning precision is difficult to meet engineering requirements. Therefore, development of a method for detecting damage to a nonmetallic pipeline, which can overcome the defects, is needed to realize high-precision positioning, intelligent identification and distributed collaborative monitoring. Disclosure of Invention In order to overcome the above-mentioned drawbacks of the prior art, embodiments of the present invention provide an acoustic emission nonmetallic pipeline damage detection method based on an intelligent sensor, so as to solve the problems set forth in the above-mentioned background art. In order to achieve the above purpose, the present invention provides the following technical solutions: A1, arranging a plurality of intelligent acoustic emission sensor nodes on the outer wall of a nonmetallic pipeline at equal intervals along the axial direction, wherein a piezoelectric film sensor, a temperature compensation module, an edge calculation unit and a wireless communication module are arranged in each node, microsecond synchronization is realized among the nodes through a time synchronization protocol, and a distributed self-networking monitoring system is formed; A2, each intelligent sensor node acquires an acoustic emission signal in real time, and after the acoustic emission signal is amplified by a preamplifier, the acoustic emission signal is subjected to self-adaptive filtering and background noise suppression by an edge computing unit, and an effective acoustic emission waveform is extracted; a3, extracting time domain characteristic parameters and frequency domain characteristic parameters of the acoustic emission signals by an edge computing unit, constructing a characteristic classification model based on a lightweight convolutional neural network, and realizing preliminary identification of damage types at a node end; a4, when a plurality of adjacent nodes detect effective acoustic emission events at the same time, each node uploads an event time stamp, a feature vector a