Method, system and medium for detecting corrosion of aluminum sheath using variational autoencoding network

A variational self-encoding network with discrete wavelet transforms and enhanced layers addresses the challenge of detecting aluminum sheath corrosion in high-voltage cables, enhancing detection accuracy and reducing noise interference.

JP7588691B2Active Publication Date: 2024-11-22GUANGZHOU POWER SUPPLY BUREAU GUANGDONG POWER GRID CO LTD
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
JP2023134215
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Priority Date
2022-11-07
Filing Date
2023-08-21
Publication Date
2024-11-22
Estimated Expiration
2043-08-21

AI Technical Summary

Technical Problem

Conventional methods struggle to accurately detect corrosion in aluminum sheaths of high-voltage cables due to multimodal frequency dispersion and noise interference in waveguide signals, limiting the effectiveness of data-driven algorithms in capturing complex relationships and extracting feature information.

Method used

A variational self-encoding network is employed to process axial and circumferential waveguide signals, utilizing discrete wavelet transforms to generate time-frequency diagrams, and construct models for corrosion detection, incorporating multiple convolutional layers and Dropout methods to enhance feature learning and generalization.

Benefits of technology

The method improves the accuracy and precision of corrosion detection in aluminum sheaths by effectively extracting features and reducing noise interference, enabling timely maintenance and reducing the risk of cable defects.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 0007588691000013
    Figure 0007588691000013
  • Figure 0007588691000014
    Figure 0007588691000014
  • Figure 0007588691000015
    Figure 0007588691000015
Patent Text Reader

Abstract

To provide a method for detecting corrosion of an aluminum sheath using a variational self-coding network.SOLUTION: The present invention relates to the technical field of non-destructive detection, in particular to a method, system and medium for detecting corrosion of an aluminum sheath using a variational self-coding network. The method provided by the present invention acquires axial and circumferential waveguide signals, acquires a time-frequency diagram of the axial and circumferential waveguide signals by using discrete wavelet transformation, and constructs a cable aluminum sheath corrosion detection variational self-coding network model to position defects in axial and circumferential directions by using the cable aluminum sheath corrosion detection variational self-coding network model, thereby realizing corrosion detection of a cable aluminum sheath with high accuracy.SELECTED DRAWING: Figure 1
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present invention relates to the technical field of non-destructive detection, and in particular to a method, system and medium for detecting aluminum sheath corrosion using a variational auto-encoding network. [Background technology]

[0002] High voltage cables are an important component of the power system and the part with the highest demand. They are composed of four parts: conductor, insulation layer, shielding layer and protective layer. The protective layer can be divided into an external protective layer and an internal protective layer. The external protective layer is generally PVC insulating material, and the internal protective layer is generally a metal layer, and usually uses a corrugated aluminum sheath. During the installation process and long-term exposure to complex engineering, the waterproof ability of the cable is greatly reduced, and electrochemical corrosion occurs during operation, causing corrosion and defects on the surface of the corrugated aluminum sheath, and in severe cases, the cable will be short-circuited and cause a major accident. The aluminum sheath is wrapped in an insulating layer, and traditional detection methods cannot detect the corrosion and damage of the aluminum sheath, so introducing ultrasonic guided wave detection technology to detect the damage of the cable aluminum sheath is a popular direction in the prior art. Summary of the Invention [Problem to be solved by the invention]

[0003] With the rapid development of artificial intelligence, methods such as classifiers, neural networks, etc. have been widely applied. However, due to the multi-modal and frequency dispersion characteristics of guided waves and the interference of noise on guided waves, traditional data-driven algorithms cannot capture the complex relationship between ultrasonic guided waves and key features. To solve this challenge, deep learning algorithms with autoencoding networks have been successfully and effectively applied to various fields of computer vision, including damage recognition, due to their feature learning and global optimization capabilities. However, traditional autoencoding networks can reproduce the input by learning a code, but the data distribution in the hidden space is uneven, that is, some areas in the hidden space cannot represent any of the observed data. And because the noise of the collected ultrasonic guided waves is high, it is difficult to extract useful feature information using traditional methods.

[0004] In view of this, a first objective of the present invention is to provide a method for detecting aluminum sheath corrosion using a variational auto-encoding network, by obtaining axial and circumferential guided wave signals, obtaining time-frequency diagrams of the axial and circumferential guided wave signals by discrete wavelet transform, constructing a cable aluminum sheath corrosion detection variational auto-encoding network model, and performing axial and circumferential location of defects by the cable aluminum sheath corrosion detection variational auto-encoding network model, thereby realizing high-precision detection of cable aluminum sheath corrosion.

[0005] Based on the same inventive concept, the second object of the present invention is to provide an aluminum sheath corrosion detection system using a variational auto-encoding network.

[0006] Based on the same inventive concept, the third object of the present invention is to provide a storage medium. [Means for solving the problem]

[0007] The first object of the present invention can be achieved by the following technical solutions: A method for detecting corrosion of an aluminum sheath using a variational auto-encoding network, comprising the steps of: Constructing a cable aluminum sheath corrosion simulation sample and acquiring axial and circumferential guided wave original signals of the cable aluminum sheath corrosion simulation sample; converting the original axial and azimuthal guided signals into time-frequency diagrams to obtain an axial time-frequency diagram and a azimuthal time-frequency diagram; A step of constructing a variational auto-encoding network model for detecting corrosion of a cable aluminum sheath, the variational auto-encoding network model including an axial cable aluminum sheath corrosion detection variational auto-encoding network model and a circumferential cable aluminum sheath corrosion detection variational auto-encoding network model; inputting the axial time-frequency diagram into an axial cable aluminum sheath corrosion detection variational auto-encoding network model for training; inputting the circumferential time-frequency diagram into a circumferential cable aluminum sheath corrosion detection variational auto-encoding network model for training; Obtaining an axial guided wave original signal of the cable aluminum sheath to be detected, and performing axial positioning for the corrosion defect of the cable aluminum sheath using the trained axial cable aluminum sheath corrosion detection variational auto-encoding network model; The method includes a step of acquiring an original circumferential guided wave signal of the cable aluminum sheath to be detected, and performing circumferential positioning of the corrosion defects of the cable aluminum sheath using the trained circumferential cable aluminum sheath corrosion detection variational auto-encoding network model.

[0008] Furthermore, the step of transforming the axial and circumferential guided signals into a time-frequency diagram by a discrete wavelet transform comprises: Passing the axial original signal through a high-pass filter and a low-pass filter respectively, and decomposing it into a series of finite attenuation signals through a down-sampling filter, and calculating the axial original signal time-frequency diagram through mother wavelet calculation and signal transformation; The method includes the steps of passing the circumferential original signal through a high-pass filter and a low-pass filter, respectively, and decomposing the circumferential original signal into a series of finitely attenuated signals by a down-sampling filter, and calculating a circumferential original signal time-frequency diagram by a mother wavelet calculation and signal transformation.

[0009] Furthermore, the variational autoencoding network model for detecting corrosion of an aluminum sheath of a cable includes an encoder and a decoder, and the variational autoencoding network is trained to minimize the reconstruction error between the encoded and decoded data and the initial data, The encoder compresses the vectors in the input feature space into the hidden feature space to obtain the hidden variables, and the distribution of the hidden variables approaches a normal distribution. The decoder restores a representation of the hidden feature space to the input feature space.

[0010] Furthermore, multiple convolutional layers and pooling layers are added to the encoder part of the variational auto-encoding network model for cable aluminum sheath corrosion detection, and the encoder is constructed using the Dropout method.

[0011] Furthermore, the performance of the reconstructed signal of the variational autoencoding network model is evaluated by the ELBO loss function LOSS, whose expression is shown in the following equation 1.

[0012]

number

[0013] The second object of the present invention can be achieved by the following technical solutions: An aluminum sheath corrosion detection system using a variational autoencoding network, A sample construction module for constructing cable aluminum sheath corrosion simulation samples; A data acquisition module for acquiring original axial and circumferential guided wave signals; a time-frequency diagram conversion module for converting the original axial and circumferential guided signals into a time-frequency diagram to obtain an axial time-frequency diagram and a circumferential time-frequency diagram; A model construction module for constructing a variational auto-encoding network model for detecting corrosion of a cable aluminum sheath, the model construction module including an axial-direction cable aluminum sheath corrosion detection variational auto-encoding network model and a circumferential-direction cable aluminum sheath corrosion detection variational auto-encoding network model; a model training module for inputting the axial time-frequency diagram into an axial cable aluminum sheath corrosion detection variational auto-encoding network model for training, and inputting the circumferential time-frequency diagram into a circumferential cable aluminum sheath corrosion detection variational auto-encoding network model for training; and a model recognition module that uses the trained axial cable aluminum sheath corrosion detection variational auto-encoding network model to perform axial positioning for corrosion defects in the cable aluminum sheath, and uses the trained circumferential cable aluminum sheath corrosion detection variational auto-encoding network model to perform circumferential positioning for corrosion defects in the cable aluminum sheath.

[0014] Furthermore, the data acquisition module includes a computer, a signal generator, a signal amplifier, a signal acquisition card, and a solid coupling medium, where the solid coupling medium matches the acoustic impedance of the object to be detected.

[0015] Furthermore, the solid bonding medium is an epoxy resin.

[0016] The third object of the present invention can be achieved by the following technical solutions: A storage medium on which a program is stored, the program realizing the above-mentioned aluminum sheath corrosion detection method using a variational auto-encoding network when executed by a computer. Effect of the Invention

[0017] The present invention has the following beneficial effects over the prior art: (1) In the present invention, axial and circumferential guided wave signals are obtained, and time-frequency diagrams of the axial and circumferential guided wave signals are obtained by discrete wavelet transform. A cable aluminum sheath corrosion detection variational auto-encoding network model is constructed, and the axial and circumferential locations of defects are determined by the cable aluminum sheath corrosion detection variational auto-encoding network model, thereby realizing the extraction of effective feature information from the signals, improving the accuracy of the detection and location of the cable aluminum sheath, and maintaining the cable aluminum sheath in a timely manner. (2) The variational auto-encoding network model provided by the present invention adds multiple convolutional layers and pooling layers to the encoder part, and constructs the encoder using the Dropout method, which can improve the generalization ability of the model and ensure that the hidden space has good attributes for carrying out the data generation process. (3) The distribution of the hidden variables of the autoencoding network of the present invention approaches the standard normal distribution, so that the hidden variables have stable statistical properties, which improves the convenience and applicability of the decoder. (4) In the present invention, the ultrasonic signal excited by the solid-state coupling method can smoothly enter the object to be detected, and the influence of noise is further reduced in the aluminum sheath corrosion detection method using the variational auto-encoding network provided by the present invention. [Brief description of the drawings]

[0018] [Figure 1] 1 is a flowchart of a cable aluminum sheath corrosion detection method using a variational auto-encoding network according to a first embodiment of the present invention. [Diagram 2] 1 is a flowchart of a discrete wavelet transform according to a first embodiment of the present invention. [Diagram 3] FIG. 1 is a structural diagram of a variational autoencoding network model according to a first embodiment of the present invention. [Figure 4] FIG. 2 is a schematic diagram of the internal operation of the variational auto-encoding network of the first embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0019] In order to make the objectives, technical solutions and advantages of the embodiments of the present invention more comprehensible, the technical solutions of the embodiments of the present invention will be described clearly and completely below with reference to the drawings of the embodiments of the present invention. It is clear that the described embodiments are not all the embodiments, but are only some of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments within the scope that can be easily conceived by those skilled in the art belong to the protection scope of the present invention.

[0020] Example 1: As shown in FIG. 1, this embodiment provides a cable aluminum sheath corrosion detection method based on variational auto-encoding network, which includes the following steps: S10: construct a cable aluminum sheath corrosion simulation sample, and obtain the axial and circumferential guided wave original signals of the cable aluminum sheath corrosion simulation sample; In this embodiment, the cable aluminum sheath corrosion defect samples include uniform corrosion, pitting corrosion and filiform corrosion, and are obtained by artificially using heavy blocks to simulate or collect real corrosion defect samples; S20: Convert the original axial and circumferential guided signals into time-frequency diagrams; In this embodiment, the step of transforming the axial and circumferential guided signals into a time-frequency diagram by discrete wavelet transform includes the following steps: S21: The axial original signal is passed through a high-pass filter and a low-pass filter respectively, and is decomposed into a series of finite attenuation signals by a down-sampling filter, and the axial original signal time-frequency diagram is calculated by mother wavelet calculation and signal transformation; S22: The original circumferential signal is passed through a high-pass filter and a low-pass filter respectively, and is decomposed into a series of finitely attenuated signals by a down-sampling filter, and the time-frequency diagram of the original circumferential signal is calculated by mother wavelet calculation and signal transformation; In steps S210 and S220 of this embodiment, the discrete scale parameter a and the translation parameter b ensure that the signal time-frequency diagram can be obtained with limited calculation. a and b are discretized by the following formula 2.

[0021]

number

[0022] In the discretization process, a 0 =2 and b 0 Assuming that =1, the mother wavelet is expressed as follows:

[0023]

number

[0024] In this case, the discrete wavelet transform is expressed as in the following equation 4.

[0025]

number

[0026] The discretization is with respect to the scale parameter a and the translation parameter b, and within one space the time variable t may still be defined as continuous.

[0027] S30: A variational auto-encoding network model for detecting corrosion of a cable aluminum sheath is constructed, including a variational auto-encoding network model for detecting corrosion of an axial cable aluminum sheath and a variational auto-encoding network model for detecting corrosion of a circumferential cable aluminum sheath.

[0028] In this embodiment, the variational autoencoding network model for detecting corrosion of an aluminum sheath of a cable includes an encoder and a decoder, and the variational autoencoding network is trained to minimize the reconstruction error between the encoded and decoded data and the initial data; The encoder compresses the vectors in the input feature space into the hidden feature space to obtain the hidden variables, and the distribution of the hidden variables approaches a normal distribution. The encoder part is constructed using the Dropout method by adding multiple convolutional layers and pooling layers.

[0029] The decoder restores a representation of the hidden feature space to the input feature space.

[0030] In this embodiment, the internal operation of the variational autoencoding network model is as follows: In this embodiment, the performance of the reconstructed signal of the variational autoencoding network model is evaluated by the ELBO loss function LOSS, which is calculated as follows: The encoder part of the whole neural network is designed to represent a conditional probability q(z|x), where x is the time-frequency diagram data input to the model and z is a hidden variable. Then the decoder tries to recover the original data using the hidden variable z. Specifically, one neural network represents one conditional probability p(x|z).

[0031] First, we introduce a hidden variable z, and then denote p(x) in terms of z:

[0032]

number

[0033] Autoencoding networks have the same encoding and decoding process, and should have the same output as much as possible for the same input. In the above formula, z represents p(x), so for any input data, there is a hidden variable at the end, which should ensure that the converted output data and the input data are as equal as possible.

[0034] By applying a transformation to log p(x):

[0035]

number

[0036] q(z|x) is an arbitrary distribution.

[0037] To incorporate the encoding process, we introduce q(z|x) to obtain the following equation (7).

[0038]

number

[0039] By deformation, L b The expression is obtained as shown in the following formula 8.

[0040]

number

[0041] To maximize p(x), find p(x|z) and q(z|x) and b Make as large as possible; By adjusting q(z|x), the KL value is minimized and the value of number 9 in the above equation is maximized.

[0042]

number

[0043] After parameter tuning using the neural network, KL(q(z|x)||p(z|x))=0, and L b is considered as the ELBO, as shown in the following formula:

[0044]

number

[0045] Finally, the expression for the ELBO loss function LOSS is as follows:

[0046]

number

[0047] S40: Input the axial time-frequency diagram into the axial cable aluminum sheath corrosion detection variational auto-encoding network model and train it; S50: The circumferential time-frequency diagram is input into the circumferential cable aluminum sheath corrosion detection variational auto-encoding network model for training; S60: Obtain the axial guided wave original signal of the cable aluminum sheath to be detected, and use the trained axial cable aluminum sheath corrosion detection variational auto-encoding network model to perform axial positioning for the corrosion defect of the cable aluminum sheath.

[0048] In this embodiment, step S60 includes the following steps: S61: obtain an original axial guided wave signal of the aluminum sheath of the cable to be detected; S62: Convert the axial guided wave signal into a time-frequency diagram by discrete wavelet transform; S63: The axial time-frequency diagram is input into the axial cable aluminum sheath corrosion detection variational auto-encoding network model trained in step S40 to perform feature recognition to realize the axial location of the corrosion defects in the cable aluminum sheath. S70: Obtain the circumferential guided wave original signal of the cable aluminum sheath to be detected, and use the trained circumferential cable aluminum sheath corrosion detection variational auto-encoding network model to perform circumferential positioning for the corrosion defects of the cable aluminum sheath.

[0049] In this embodiment, step S70 includes the following steps: S71: Obtain an original circumferential guided wave signal of the aluminum sheath of the cable to be detected; S72: converting the circumferential guided wave signal into a time-frequency diagram by discrete wavelet transform; S73: The circumferential time-frequency diagram is input into the circumferential cable aluminum sheath corrosion detection variational auto-encoding network model trained in step S50 to perform feature recognition, thereby realizing the circumferential location of the corrosion defects of the cable aluminum sheath.

[0050] As described above, in this embodiment, the axial and circumferential guided wave signals are obtained, and the time-frequency diagrams of the axial and circumferential guided wave signals are obtained by discrete wavelet transform. A variational auto-encoding network model for cable aluminum sheath corrosion detection is constructed, and the axial and circumferential location of defects is performed by the variational auto-encoding network model for cable aluminum sheath corrosion detection, thereby realizing the extraction of effective feature information from the signal, improving the accuracy of the detection and location of the cable aluminum sheath, and maintaining the cable aluminum sheath in a timely manner. The variational auto-encoding network model provided in this embodiment adds multiple convolution layers and pooling layers to the encoder part, and the encoder is constructed by the Dropout method, which can improve the generalization ability of the model, and ensures that the hidden space has good attributes for performing the data generation process. The distribution of the hidden variables of the auto-encoding network of this embodiment is close to a standard normal distribution, so that the hidden variables have stable statistical properties, and improve the convenience and application rate of the decoder.

[0051] Example 2: This embodiment provides an aluminum sheath corrosion detection system using a variational autoencoding network, A sample construction module for constructing cable aluminum sheath corrosion simulation samples; A data acquisition module for acquiring original axial and circumferential guided wave signals; a time-frequency diagram conversion module for converting the original axial and circumferential guided signals into a time-frequency diagram to obtain an axial time-frequency diagram and a circumferential time-frequency diagram; A model construction module for constructing a variational auto-encoding network model for detecting corrosion of a cable aluminum sheath, the model construction module including an axial-direction cable aluminum sheath corrosion detection variational auto-encoding network model and a circumferential-direction cable aluminum sheath corrosion detection variational auto-encoding network model; a model training module for inputting the axial time-frequency diagram into an axial cable aluminum sheath corrosion detection variational auto-encoding network model for training, and inputting the circumferential time-frequency diagram into a circumferential cable aluminum sheath corrosion detection variational auto-encoding network model for training; and a model recognition module that uses the trained axial cable aluminum sheath corrosion detection variational auto-encoding network model to perform axial positioning for corrosion defects in the cable aluminum sheath, and uses the trained circumferential cable aluminum sheath corrosion detection variational auto-encoding network model to perform circumferential positioning for corrosion defects in the cable aluminum sheath.

[0052] In this embodiment, the data acquisition module includes a computer, a signal generator, a signal amplifier, a signal acquisition card, and a solid coupling medium, where the solid coupling medium matches the acoustic impedance of the object to be detected.

[0053] In this example, the solid bonding medium is an epoxy resin.

[0054] That is, in the above modules of this embodiment, the sample construction module and the data acquisition module cooperate to realize step S10 of the first embodiment of the present invention, the time-frequency diagram conversion module realizes step S20 of the first embodiment, the model construction module realizes step S30 of the first embodiment, the model training module realizes steps S40 and S50 of the first embodiment, and the model recognition module and the data acquisition module cooperate to realize steps S60 and S70 of the first embodiment. Since steps S10 to S70 have been described in detail in the first embodiment, in order to simplify the description of the specification, the detailed realization process of each of the above modules in this embodiment can be referred to in the first embodiment, and no further description will be given.

[0055] In this embodiment, the ultrasonic signal excited in the solid-state coupling manner can smoothly enter the object to be detected, and the influence of noise is further reduced in the aluminum sheath corrosion detection method using the variational auto-encoding network provided by the present invention.

[0056] Example 3: This embodiment provides a storage medium on which a program is stored, and when the program is executed by a computer, the method for detecting aluminum sheath corrosion using a variational auto-encoding network according to the embodiment of the present invention is realized; Constructing a cable aluminum sheath corrosion simulation sample and acquiring axial and circumferential guided wave original signals of the cable aluminum sheath corrosion simulation sample; converting the original axial and azimuthal guided signals into time-frequency diagrams to obtain an axial time-frequency diagram and a azimuthal time-frequency diagram; A step of constructing a variational auto-encoding network model for detecting corrosion of a cable aluminum sheath, the variational auto-encoding network model including an axial cable aluminum sheath corrosion detection variational auto-encoding network model and a circumferential cable aluminum sheath corrosion detection variational auto-encoding network model; inputting the axial time-frequency diagram into an axial cable aluminum sheath corrosion detection variational auto-encoding network model for training; inputting the circumferential time-frequency diagram into a circumferential cable aluminum sheath corrosion detection variational auto-encoding network model for training; Obtaining an axial guided wave original signal of the cable aluminum sheath to be detected, and performing axial positioning for the corrosion defect of the cable aluminum sheath using the trained axial cable aluminum sheath corrosion detection variational auto-encoding network model; The method includes a step of acquiring an original circumferential guided wave signal of the cable aluminum sheath to be detected, and performing circumferential positioning of the corrosion defects of the cable aluminum sheath using the trained circumferential cable aluminum sheath corrosion detection variational auto-encoding network model.

[0057] Here, the computer-readable storage medium of this embodiment may be a computer-readable signal medium, or a computer-readable storage medium, or any combination of the above. The computer-readable storage medium may be, for example, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of the computer-readable storage medium may include, but are not limited to, an electrical connection having one or more leads, a portable computer magnetic disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a compact disk read-only memory (CD-ROM), an optical memory, a magnetic memory, or any suitable combination of the above.

[0058] In this embodiment, the computer-readable storage medium may be any tangible medium that contains or stores a program that is used by or in combination with an instruction execution system, apparatus, or device. In this embodiment, the computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier, and a computer-readable program is carried on it. Such a propagated data signal may take many forms, including, but not limited to, an electromagnetic signal, an optical signal, or any suitable combination of the above. The computer-readable signal medium may also be any computer-readable storage medium other than a computer-readable storage medium, which may transmit, propagate, or transmit a program that is used by or in combination with an instruction execution system, apparatus, or device. The computer program contained in the computer-readable storage medium may be transmitted by any suitable medium, including, but not limited to, a wire, an optical cable, RF (radio frequency), or any suitable combination of the above.

[0059] The computer readable storage medium may be written in one or more program design languages, or a combination thereof, for implementing the present embodiment. The program design language may be an object-oriented program design language, such as Java. (Registered Trademark) , Smalltalk (Registered Trademark) , C++, conventional procedural program design languages ​​such as C or similar program design languages. The programs may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on the remote computer, or entirely on a remote computer or server. When referring to remote computers, the remote computers may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or connected to an external computer (e.g., connected via the Internet using an Internet Service Provider).

[0060] The above-described embodiments are not all embodiments but are merely some embodiments of the present invention, and the present invention is not limited to the details of the above embodiments, and any appropriate changes or modifications completed by those skilled in the art do not depart from the patent scope of the present invention.

Claims

1. A method for detecting corrosion of an aluminum sheath using a variational auto-encoding network, comprising the steps of: Constructing a cable aluminum sheath corrosion simulation sample and acquiring axial and circumferential guided wave original signals of the cable aluminum sheath corrosion simulation sample; converting the original axial and azimuthal guided signals into time-frequency diagrams to obtain an axial time-frequency diagram and a azimuthal time-frequency diagram; A step of constructing a variational auto-encoding network model for detecting corrosion of a cable aluminum sheath, the variational auto-encoding network model including an axial cable aluminum sheath corrosion detection variational auto-encoding network model and a circumferential cable aluminum sheath corrosion detection variational auto-encoding network model; Inputting the axial time-frequency diagram into an axial cable aluminum sheath corrosion detection variational auto-encoding network model for training; inputting the circumferential time-frequency diagram into a circumferential cable aluminum sheath corrosion detection variational auto-encoding network model for training; Obtaining an original axial guided wave signal of the cable aluminum sheath to be detected, and inputting a time-frequency diagram converted from the original axial guided wave signal of the cable aluminum sheath to be detected into the trained axial cable aluminum sheath corrosion detection variational auto-encoding network model to perform axial positioning for the corrosion defect of the cable aluminum sheath; a step of acquiring an original circumferential guided wave signal of the cable aluminum sheath to be detected, inputting a time-frequency diagram converted from the original circumferential guided wave signal of the cable aluminum sheath to be detected into a trained circumferential cable aluminum sheath corrosion detection variational auto-encoding network model, and performing circumferential positioning of corrosion defects in the cable aluminum sheath, the step comprising:

2. The method for detecting aluminum sheath corrosion using a variational auto-encoding network as claimed in claim 1, characterized in that the step of converting the axial and circumferential guided wave signals into a time-frequency diagram specifically comprises converting the axial and circumferential guided wave signals into a time-frequency diagram by a discrete wavelet transform.

3. Transforming the axial and circumferential guided signals into a time-frequency diagram by a discrete wavelet transform, comprising: Passing the axial original signal through a high-pass filter and a low-pass filter respectively, and decomposing it into a series of finite attenuation signals through a down-sampling filter, and calculating the axial original signal time-frequency diagram through mother wavelet calculation and signal transformation; The method for detecting aluminum sheath corrosion using a variational auto-encoding network according to claim 2, further comprising the steps of: passing the circumferential original signal through a high-pass filter and a low-pass filter, respectively, and decomposing the circumferential original signal into a series of finitely attenuated signals by a down-sampling filter; and calculating a time-frequency diagram of the circumferential original signal through mother wavelet calculation and signal transformation.

4. The variational autoencoding network model for detecting corrosion of an aluminum sheath of a cable includes an encoder and a decoder, and the variational autoencoding network is trained to minimize the reconstruction error between the encoded and decoded data and the initial data; The encoder compresses the vectors in the input feature space into the hidden feature space to obtain the hidden variables, and the distribution of the hidden variables approaches a normal distribution. The method for detecting corrosion of an aluminum sheath using a variational auto-encoding network according to claim 1, characterized in that the decoder restores a representation of the hidden feature space to the input feature space.

5. The method for detecting corrosion of an aluminum sheath using a variational auto-encoding network according to claim 4, characterized in that the encoder part of the variational auto-encoding network model for detecting corrosion of an aluminum sheath of a cable is constructed using the Dropout method by adding multiple convolutional layers and pooling layers to the encoder part of the variational auto-encoding network model for detecting corrosion of an aluminum sheath of a cable.

6. The method for detecting aluminum sheath corrosion using a variational auto-encoding network according to claim 4, characterized in that the performance of the reconstructed signal of the variational auto-encoding network model is evaluated by the ELBO loss function LOSS, and its expression is expressed by the following equation (1). [0010]

7. An aluminum sheath corrosion detection system using a variational autoencoding network, A sample construction module for constructing cable aluminum sheath corrosion simulation samples; A data acquisition module for acquiring original axial and circumferential guided wave signals; a time-frequency diagram conversion module for converting the original axial and circumferential guided signals into a time-frequency diagram to obtain an axial time-frequency diagram and a circumferential time-frequency diagram; A model construction module for constructing a variational auto-encoding network model for detecting corrosion of a cable aluminum sheath, the model construction module including an axial-direction cable aluminum sheath corrosion detection variational auto-encoding network model and a circumferential-direction cable aluminum sheath corrosion detection variational auto-encoding network model; a model training module for inputting the axial time-frequency diagram into an axial cable aluminum sheath corrosion detection variational auto-encoding network model for training, and inputting the circumferential time-frequency diagram into a circumferential cable aluminum sheath corrosion detection variational auto-encoding network model for training; a model recognition module that inputs a time-frequency map converted from an original axial guided wave signal of a cable aluminum sheath into a trained axial cable aluminum sheath corrosion detection variational auto-encoding network model to perform axial positioning of a corrosion defect in the cable aluminum sheath, and inputs a time-frequency map converted from an original circumferential guided wave signal of the cable aluminum sheath into a trained circumferential cable aluminum sheath corrosion detection variational auto-encoding network model to perform circumferential positioning of a corrosion defect in the cable aluminum sheath.

8. The aluminum sheath corrosion detection system using variational self-encoding network as described in claim 7, characterized in that the data acquisition module includes a computer, a signal generator, a signal amplifier, a signal capture card, and a solid coupling medium, and the solid coupling medium matches the acoustic impedance of the object to be detected.

9. The aluminum sheath corrosion detection system using a variational auto-encoding network as described in claim 8, characterized in that the solid bonding medium is epoxy resin.

10. A storage medium on which a program is stored, the storage medium being characterized in that, when the program is executed by a computer, the storage medium realizes the aluminum sheath corrosion detection method using a variational auto-encoding network according to any one of claims 1 to 6.

Citation Information

Patent Citations

  • Buried pipeline subway stray current corrosion protection method

    CN110750880A

  • Pipeline anti-corrosion layer tiny damage point positioning method based on stationary wavelet transform

    CN112899692A

  • Corrosion occurrence detection method, corrosion occurrence detection device and corrosion occurrence detection system

    JP2012127774A