Data-driven high-voltage cable aluminum sheath defect identification method, system and medium

Abstract

To provide a data-driven high-voltage cable aluminum sheath defect identification method.SOLUTION: By introducing a dimension reduction matrix, an ultrasonic guided wave original signal is reconstructed to realize data dimension reduction, redundant information in the ultrasonic guided wave original signal is reduced, and meanwhile, the loss of related information of corrosion defects in the ultrasonic guided wave original signal is lessened, and a neural network model based on defect identification of the cable aluminum sheath is constructed to identify the defects of the cable aluminum sheath, so that the calculation accuracy and the calculation efficiency of the defect identification of the high-voltage cable aluminum sheath are improved.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to the technical field of non-destructive detection, and in particular to a data-driven high voltage cable aluminum sheath defect recognition method, system and medium. [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 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 damage of the aluminum sheath, so the introduction of ultrasonic guided wave detection technology to detect the damage of the cable aluminum sheath has become a popular direction in the prior art. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] China Patent Application Publication No. 115310474 Summary of the Invention [Problem to be solved by the invention]

[0004] 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 with guided wave signals, guided wave data expresses diversified, large-scale, high-dimensional and other characteristics, and contains rules and information with rich content, but is often covered by a large amount of redundant data, making it difficult to intuitively observe. Therefore, traditional data-driven algorithms cannot capture the complex relationship between ultrasonic guided wave signals and key features, resulting in low recognition accuracy, excessive computational complexity and other problems.

[0005] In view of this, the first objective of the present invention is to provide a data-driven method for recognizing defects in aluminum sheaths of high-voltage cables, by introducing a dimension reduction matrix to reconstruct the original ultrasonic guided wave signal to reduce the dimension of data, thereby reducing the redundant information in the original ultrasonic guided wave signal and reducing the related information loss of corrosion defects in the original ultrasonic guided wave signal, so as to construct a neural network model for recognizing defects in aluminum sheaths of high-voltage cables to recognize the defects, and to improve the calculation accuracy and calculation efficiency of the defects in aluminum sheaths of high-voltage cables.

[0006] Based on the same inventive concept, the second object of the present invention is to provide a data-driven high voltage cable aluminum sheath defect recognition system.

[0007] Based on the same inventive concept, the third object of the present invention is to provide a storage medium. [Means for solving the problem]

[0008] The first object of the present invention can be achieved by the following technical solutions: A data-driven high voltage cable aluminum sheath defect recognition method, comprising: A step of constructing a cable aluminum sheath corrosion simulating sample and acquiring an ultrasonic guided wave original signal of the cable aluminum sheath corrosion simulating sample; Pre-processing the collected ultrasonic guided wave original signals to generate training samples and test samples; merging the training samples and the test samples and reconstructing the training samples and the test samples with reduced dimensionality; Building a cable aluminum sheath defect recognition neural network model, and training the cable aluminum sheath defect recognition neural network model using the reconstructed training samples and test samples; and recognizing corrosion defects in the cable aluminum sheath using a cable aluminum sheath defect recognition neural network model.

[0009] In addition, the cable aluminum sheath corrosion defect samples include uniform corrosion, pitting corrosion and filiform corrosion defect samples, which are obtained by artificially using heavy blocks to simulate or collect real corrosion defect samples.

[0010] Furthermore, the step of performing pre-processing on the collected ultrasonic guided wave original signals to generate training samples and test samples includes: A step of performing a normalization process on the amplitude value scale of each signal, so as to make the amplitude values ​​of the collected guided signals have the same numerical scale, and to avoid the signal reconstruction network from biasedly extracting signals with large amplitude value features, the normalization expression is as follows:

number

number

number

[0011] Furthermore, the step of merging the training samples and the test samples and reconstructing the training samples and the test samples by reducing the number of dimensions of the training samples and the test samples may include: Calculating a distance matrix to fuse the time-frequency information of the training sample and test sample guided signals; constructing an optimization problem target function to solve the dimensionality reduction matrix; and reconstructing the training and test samples using the dimensionality reduction matrix.

[0012] In addition, a distance matrix is ​​calculated to fuse the time-frequency information of the training sample and the test sample guided signal, and the distance matrix

number

number

number

number

number

number

number

number

[0013] Furthermore, the expression of the optimization problem is as follows:

number

number

number

number

number

[0014] Furthermore, the cable aluminum sheath defect recognition neural network model is a recurrent neural network model based on GRU or a recurrent neural network model based on LSTM.

[0015] Furthermore, the cable aluminum sheath defect recognition neural network model is a recurrent neural network model based on GRU.

[0016] The second object of the present invention can be achieved by the following technical solutions: A data-driven high voltage cable aluminum sheath defect recognition system, comprising: A sample construction module for constructing a cable aluminum sheath corrosion simulation sample; a data acquisition module for acquiring an ultrasonic guided wave original signal; A pre-processing module performs pre-processing on the collected ultrasonic guided wave original signals to generate training samples and test samples; a dimensionality reduction and reconstruction module that merges the training samples and the test samples and reduces the dimensionality of the training samples and the test samples to reconstruct them; a model training module for constructing a neural network model for recognizing defects in a cable aluminum sheath and training the neural network model for recognizing defects in a cable aluminum sheath using the reconstructed training samples and the test samples; and a model recognition module for recognizing corrosion defects in the cable aluminum sheath using a cable aluminum sheath defect recognition neural network model.

[0017] 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 implementing the above-mentioned data-driven high-voltage cable aluminum sheath defect recognition method when executed by a computer. Effect of the Invention

[0018] The present invention has the following beneficial effects over the prior art: (1) In the present invention, a dimensionality reduction matrix is ​​introduced to reconstruct the original ultrasonic guided wave signal, thereby realizing data dimensionality reduction, reducing redundant information in the original ultrasonic guided wave signal, and improving the calculation efficiency of high-voltage cable aluminum sheath defect recognition. (2) In the present invention, the training samples and the test samples are merged, and the dimensionality of the training samples and the test samples is reduced and reconstructed. According to the reconstruction method of the present invention, the redundant information in the ultrasonic guided wave original signal is reduced, and the loss of relevant information of corrosion defects in the ultrasonic guided wave original signal is reduced, thereby improving the calculation accuracy of high-voltage cable aluminum sheath defect recognition. (3) In the present invention, a GRU-based recurrent neural network model is constructed, and the GRU-based recurrent neural network model is used to train and recognize the reconstruction samples, thereby solving problems such as long-term memory and gradations in error backpropagation, and further improving the calculation accuracy and calculation efficiency of high-voltage cable aluminum sheath defect recognition, thereby enhancing the reliability and security of power system operation. [Brief description of the drawings]

[0019] In order to more clearly explain the technical solutions of the embodiments of the present application, the following briefly introduces the drawings required in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application, and those skilled in the art can obtain other drawings based on these drawings on the premise that they do not make efforts worthy of inventive step. [Figure 1] 1 is a flowchart of a data-driven high-voltage cable aluminum sheath defect recognition method according to a first embodiment of the present invention; [Diagram 2] 1 is a flowchart of network reconfiguration in the data-driven high-voltage cable aluminum sheath defect recognition method according to embodiment 1 of the present invention; [Diagram 3] FIG. 1 is a structural diagram of a GRU input / output according to the first embodiment of the present invention. [Figure 4] FIG. 2 is a diagram showing the internal structure of a GRU according to the first embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0020] In order to make the objectives, technical solutions and advantages of the embodiments of the present invention more comprehensible, the following clearly and completely describes the technical solutions of the embodiments of the present invention in combination with the drawings of the embodiments of the present invention. Obviously, the described embodiments are not all the embodiments, but only some of the embodiments of the present invention. Based on the embodiments of the present invention, on the premise that a person skilled in the art does not perform any work worthy of inventive step, all other embodiments obtained fall within the protection scope of the present invention.

[0021] Example 1: As shown in FIG. 1, this embodiment provides a data-driven high voltage cable aluminum sheath defect recognition method, which includes the following steps: S10: Construct a cable aluminum sheath corrosion simulating sample, and obtain an ultrasonic guided wave original signal of the cable aluminum sheath corrosion simulating sample; In this embodiment, the cable aluminum sheath corrosion defect samples include uniform corrosion, pitting corrosion and filiform corrosion defect samples, which are obtained by artificially using heavy blocks to simulate or collect real corrosion defect samples.

[0022] S20: The step of preprocessing the collected ultrasonic guided wave original signal to generate training samples and test samples is specifically as follows: In order to make the amplitude values ​​of the collected guided signals have the same numerical scale, and to avoid the signal reconstruction network from biasedly extracting signals with large amplitude value characteristics, a normalization process is performed on the amplitude value scale of each signal, and the normalization expression is as follows:

number

number

number

[0023] S30: Merge the training samples and the test samples, and reduce the dimensionality of the training samples and the test samples to reconstruct them, as shown in FIG. 2, including the following steps: S31: Calculate a distance matrix to fuse the time-frequency information of the training sample and the test sample guided wave signals; In this embodiment, the distance matrix in step S31

number

number

number

number

number

number

number

number

[0024] S32: Construct the optimization problem target function and solve the dimensionality reduction matrix; In this embodiment, the expression of the optimization problem is as follows:

number

number

number

number

number

[0025] S33: Reconstruct the training samples and test samples using the dimensionality reduction matrix, specifically, PY is the reconstructed test sample, and PX is the reconstructed training sample.

[0026] S40: Construct a cable aluminum sheath defect recognition neural network model, and train the cable aluminum sheath defect recognition neural network model using the reconstructed training sample and the test sample; As shown in FIG. 3 and FIG. 4, in this embodiment, the cable aluminum sheath defect recognition neural network model is a recurrent neural network model using GRU, and the input of GRU is x at time t. t and the hidden layer state h at time t-1 t-1 The hidden state contains the relevant information of the previous node, and the output of the GRU is the output y of the hidden node at time t. t and the hidden state h transmitted to the next node t It consists of:

[0027] Previous transmitted state h t-1 and the current node's input x t Two gate states, the update gate and the reset gate, are obtained by

[0028] After obtaining the gate signal, first, h is reset by the reset gate. t-1 Reset h t-1 and x t Finally, we scale the data to the range of -1 to 1 using a tanh activation function.

number

number

number

number

[0029] For the current and previous hidden units, the reset gates determine how new input information is combined with the previous memory, and the update gates define how much of the previous memory is retained in the current time step.

number

number

[0030] In another embodiment of the present invention, the cable aluminum sheath defect recognition neural network model is a recurrent neural network model based on LSTM.

[0031] S50: The step of recognizing corrosion defects in a cable aluminum sheath using a cable aluminum sheath defect recognition neural network model includes the following steps: S51: Cable aluminum sheath ultrasonic guided wave to obtain the original signal; S52: Pre-processing the collected ultrasonic guided wave original signal; S53: Using the dimension reduction matrix obtained in step S32, perform dimension reduction and reconstruction on the ultrasonic guided wave original signal; S54: The reconstructed signal obtained in step S53 is input into the cable aluminum sheath defect recognition neural network model obtained in step S40 to obtain the corrosion defect recognition result of the cable aluminum sheath.

[0032] As described above, in this embodiment, a dimension reduction matrix is ​​introduced to reconstruct the original ultrasonic guided wave signal to realize data dimension reduction, reduce the redundant information in the original ultrasonic guided wave signal, and improve the calculation efficiency of high-voltage cable aluminum sheath defect recognition; in this embodiment, the training sample and the test sample are merged, and the dimensionality of the training sample and the test sample are reduced and reconstructed. According to the reconstruction method of the present invention, the redundant information in the original ultrasonic guided wave signal is reduced, and the loss of related information of corrosion defects in the original ultrasonic guided wave signal is reduced, so as to improve the calculation accuracy of high-voltage cable aluminum sheath defect recognition; in this embodiment, a GRU-based recurrent neural network model is constructed, and the GRU-based recurrent neural network model is used to train and recognize the reconstructed sample, so as to further improve the calculation accuracy and calculation efficiency of high-voltage cable aluminum sheath defect recognition, and enhance the reliability and security of power system operation.

[0033] Example 2: This embodiment provides a data-driven high voltage cable aluminum sheath defect recognition system, A sample construction module for constructing a cable aluminum sheath corrosion simulation sample; a data acquisition module for acquiring an ultrasonic guided wave original signal; A pre-processing module performs pre-processing on the collected ultrasonic guided wave original signals to generate training samples and test samples; a dimensionality reduction and reconstruction module that merges the training samples and the test samples and reduces the dimensionality of the training samples and the test samples to reconstruct them; a model training module for constructing a neural network model for recognizing defects in a cable aluminum sheath and training the neural network model for recognizing defects in a cable aluminum sheath using the reconstructed training samples and the test samples; and a model recognition module for recognizing corrosion defects in the cable aluminum sheath using a cable aluminum sheath defect recognition neural network model.

[0034] 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 pre-processing module realizes step S20 of the first embodiment, the dimensionality reduction and reconstruction module realizes step S30 of the first embodiment, the model training module realizes step S40 of the first embodiment, and the model recognition module and the data acquisition module cooperate to realize step S50 of the first embodiment. Since the first embodiment describes steps S10 to S50 in detail, 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 the first embodiment, and will not be described in detail.

[0035] Example 3: A storage medium storing a program, the program being executed by a computer to realize a data-driven high-voltage cable aluminum sheath defect recognition method according to a first embodiment of the present invention; A step of constructing a cable aluminum sheath corrosion simulating sample and acquiring an ultrasonic guided wave original signal of the cable aluminum sheath corrosion simulating sample; Pre-processing the collected ultrasonic guided wave original signals to generate training samples and test samples; merging the training samples and the test samples and reconstructing the training samples and the test samples with reduced dimensionality; Building a cable aluminum sheath defect recognition neural network model, and training the cable aluminum sheath defect recognition neural network model using the reconstructed training samples and test samples; and recognizing corrosion defects in the cable aluminum sheath using a cable aluminum sheath defect recognition neural network model.

[0036] 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.

[0037] 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 conjunction 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 is loaded with a computer-readable program. 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 conjunction 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.

[0038] The computer readable storage medium may be written in one or more program design languages, or a combination thereof, for implementing the present embodiment, including object-oriented program design languages ​​such as Java, Smalltalk, C++, conventional procedural program design languages ​​such as C or similar program design languages. The program may be executed entirely on the user computer, partially on the user computer, or as a separate package, partially on the user computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user 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).

[0039] 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 data-driven high voltage cable aluminum sheath defect recognition method, comprising: A step of constructing a cable aluminum sheath corrosion simulating sample and acquiring an ultrasonic guided wave original signal of the cable aluminum sheath corrosion simulating sample; Pre-processing the collected ultrasonic guided wave original signals to generate training samples and test samples; merging the training samples and the test samples and reconstructing the training samples and the test samples with reduced dimensionality; Building a cable aluminum sheath defect recognition neural network model, and training the cable aluminum sheath defect recognition neural network model using the reconstructed training samples and test samples; and recognizing corrosion defects in the cable aluminum sheath using a cable aluminum sheath defect recognition neural network model.

2. The data-driven high-voltage cable aluminum sheath defect recognition method according to claim 1, characterized in that the cable aluminum sheath corrosion defect samples include uniform corrosion, pitting corrosion, and filiform corrosion defect samples, and are obtained by artificially using a heavy block to simulate or collect actual corrosion defect samples.

3. The step of performing pre-processing on the collected ultrasonic guided wave original signals to generate training samples and test samples includes: A normalization process is performed on the amplitude value scale of each signal, and the normalization expression is as follows: [0010] B ｉ indicates the amplitude value of the i-th signal, and B ｍａｘ indicates the maximum value of the signal, B ｍｉｎ is the step indicating the minimum value of the signal; Take N training samples, [0025] Then, we obtain N test samples, [0030] 2. The data-driven high voltage cable aluminum sheath defect recognition method according to claim 1, further comprising the steps of:

4. The step of merging the training samples and the test samples and reconstructing the training samples and the test samples with reduced dimensionality includes: Calculating a distance matrix to fuse the time-frequency information of the training sample and test sample guided signals; constructing an optimization problem target function to solve the dimensionality reduction matrix; The data-driven high-voltage cable aluminum sheath defect recognition method according to claim 3, further comprising the step of reconstructing the training samples and the test samples using the dimension reduction matrix.

5. Calculate the distance matrix by fusing the time-frequency information of the training sample and the test sample guided signal, and obtain the distance matrix [0045] is calculated using the following formula: [0050] [006] is the distance matrix [0070] The first [0080] row, no. [0090] indicates the elements of a column, [0010] denotes the amplitude difference value of the time-frequency information between different samples, ##EQU00011## 5. The data-driven high-voltage cable aluminum sheath defect recognition method according to claim 4, wherein: is a weighting parameter that controls a weight of an amplitude difference value of the time-frequency information.

6. The optimization problem is expressed as follows: ##EQU00012## P denotes the dimensionality reduction matrix, X denotes the training samples, Y denotes the test samples, and A denotes the representation coefficients of the discriminant features obtained by the matrix P; ##EQU00013## denotes the Hadamard operator, M denotes a distance matrix, J denotes auxiliary variables for constructing the distance matrix M, ##EQU00014## and ##EQU00015## are two parameters that adjust the weights of the training samples and the test samples, E is the identity matrix, F is the original dimension, and τ is a parameter that adjusts the size of the auxiliary variables. ##EQU00016## denotes the signal time-frequency map set, ##EQU00017## 6. The data-driven method for detecting defects in an aluminum sheath of a high-voltage cable according to claim 5,

7. 2. The data-driven high-voltage cable aluminum sheath defect recognition method according to claim 1, wherein the cable aluminum sheath defect recognition neural network model is a GRU-based recurrent neural network model or an LSTM-based recurrent neural network model.

8. 8. The data-driven high-voltage cable aluminum sheath defect recognition method according to claim 7, wherein the cable aluminum sheath defect recognition neural network model is a GRU-based recurrent neural network model.

9. A data-driven high voltage cable aluminum sheath defect recognition system, comprising: A sample construction module for constructing a cable aluminum sheath corrosion simulation sample; a data acquisition module for acquiring an ultrasonic guided wave original signal; A pre-processing module performs pre-processing on the collected ultrasonic guided wave original signals to generate training samples and test samples; a dimensionality reduction and reconstruction module that merges the training samples and the test samples and reduces the dimensionality of the training samples and the test samples to reconstruct them; a model training module for constructing a neural network model for recognizing defects in a cable aluminum sheath and training the neural network model for recognizing defects in a cable aluminum sheath using the reconstructed training samples and the test samples; A data-driven high-voltage cable aluminum sheath defect recognition system comprising: a model recognition module that recognizes corrosion defects in the cable aluminum sheath using a cable aluminum sheath defect recognition neural network model.

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, it realizes the data-driven high-voltage cable aluminum sheath defect recognition method described in any one of claims 1 to 8.

Images