High-voltage cable partial discharge mode identification method, system and device and storage medium

Abstract

The invention discloses a high-voltage cable partial discharge mode identification method, system and device and a storage medium, and the method comprises the following steps: obtaining partial discharge data of different high-voltage cable insulation defect types, extracting partial discharge instantaneous pulses from the partial discharge data to construct characteristic parameters of the pulses, performing normalization processing on the feature parameters to obtain original feature data; performing feature extraction on the original feature data by using SDAE, DBN and CNN to obtain deep learning high-order features; a DESCA-CatBoost classifier model is established, and the DESCA-CatBoost classifier model is trained by using the deep learning high-order features, so that the trained DESCA-CatBoost classifier model can be obtained; and the trained DESCA-CatBoost classifier is used to identify the partial discharge mode of the high-voltage cable. The method, the system, the equipment and the storage medium can accurately identify the partial discharge mode of the high-voltage cable.

Description

Technical Field

[0001] The present invention belongs to the technical field of fault detection and relates to a method, system, equipment and storage medium for identifying partial discharge patterns of high-voltage cables. Background Art

[0002] Partial discharge monitoring is an important means of assessing cable insulation condition and improving power supply reliability. Online or offline partial discharge-based condition monitoring has been applied in many industrial systems. However, because many insulation defect types vary very slightly, some partial discharge signals are highly similar and difficult to distinguish.

[0003] Scholars have conducted extensive research on partial discharge pattern recognition, with deep learning methods receiving widespread attention. However, most existing methods utilize only a single deep learning model, leaving room for improvement in recognition accuracy. Summary of the Invention

[0004] The purpose of the present invention is to overcome the shortcomings of the above-mentioned prior art and provide a method, system, device and storage medium for identifying partial discharge patterns of high-voltage cables. The method, system, device and storage medium can accurately identify partial discharge patterns of high-voltage cables.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] In one aspect, the present invention provides a method for identifying partial discharge patterns of high-voltage cables, comprising the following steps:

[0007] Acquiring partial discharge data of different high-voltage cable insulation defect types, extracting partial discharge transient pulses from the partial discharge data to construct characteristic parameters of the pulses, and normalizing the characteristic parameters to obtain original characteristic data;

[0008] Use SDAE, DBN and CNN to extract features from the original feature data to obtain deep learning high-order features;

[0009] Establishing a DESCA-CatBoost classifier model, and training the DESCA-CatBoost classifier model using the deep learning high-order features to obtain a trained DESCA-CatBoost classifier model;

[0010] The trained DESCA-CatBoost classifier is used to recognize partial discharge patterns of high-voltage cables.

[0011] The method for identifying partial discharge patterns of high-voltage cables according to the present invention is further improved in that:

[0012] The characteristic parameters of the pulse include discharge amount, peak voltage, average voltage, root mean square value, standard deviation, pulse width, rise time, fall time, equivalent time length, equivalent width, main frequency, wavelet characteristic parameters, phase angle, phase density, phase asymmetry, pulse polarity, skewness, kurtosis, crest factor, shape factor and partial discharge position.

[0013] The process of using SDAE, DBN and CNN to extract features from the original feature data obtained in step 1) to obtain deep learning high-order features is as follows:

[0014] The original feature data is input into SDAE, DBN and CNN respectively, and the features of the last hidden layer of SDAE and DBN and the output features of the last Flatten layer of CNN are extracted to construct high-order features for deep learning.

[0015] The process of constructing the DESCA-CatBoost classifier model is as follows:

[0016] The differential evolution-sine-cosine hybrid algorithm DESCA is used to optimize the hyperparameters of CatBoost, and the optimized CatBoost is used to construct the DESCA-CatBoost classifier model.

[0017] The process of optimizing the hyperparameters of CatBoost using the differential evolution-sine-cosine hybrid algorithm DESCA is as follows:

[0018] 31) Initialize the value range of each parameter of the CatBoost model and set the number of iterations of the DESCA overall optimization algorithm to T*;

[0019] 32) Randomly generate N populations, where the individuals in each population are equivalent to a set of CatBoost parameters;

[0020] 33) Using the classification accuracy as the individual fitness value, initialize the individual fitness values of all particles in the N populations in step 32);

[0021] 34) Performing a DE optimization on the N populations to obtain N optimal particles;

[0022] 35) For the fitness values of the N optimal particles, update the optimal individual representing the current iteration number according to the size of the fitness value;

[0023] 36) Return to step 34) and continue to complete DE optimization until the upper limit of the number of iterations T1 is reached. At this time, each population will have T1 historical optimal particles. The fitness of the T1 historical optimal particles is compared, and the particle with the highest fitness value is taken as the optimal individual of the population, and N optimal individuals from N populations are obtained;

[0024] 37) The N best individuals from the N populations obtained in step 36) are used as the initial population of the SCA algorithm, and SCA optimization is performed. The number of iterations of the SCA algorithm is set to T2;

[0025] 38) Initialize the parameters of the SCA algorithm, use the classification accuracy as the fitness value, calculate the fitness value of the particle, retain the optimal particle Pg, and use equations (1), (2) and (3) to update the positions of all particles:

[0026]

[0027]

[0028] in, is the jth dimension of the ith particle in the tth iteration, is the optimal particle of the tth iteration, ω is the nonlinear weight, ω max is the maximum weight, ω min is the minimum weight, r1 is a linear decreasing function, r2∈[0,2π], r3∈[-2,2], r4∈[-1,1] are random numbers, a is a constant, t is the current number of iterations, and T2 is the maximum number of iterations;

[0029] 39) Determine whether the current number of iterations is ≤ T2. If so, return to step 38); otherwise, jump to step 310);

[0030] 310) Determine whether the current total number of iterations is ≤ T*. If so, randomly select K historical optimal particles in SCA to replace the K individuals in each DE population in step 32), and return to step 32). Otherwise, output the optimal solution.

[0031] In a second aspect of the present invention, a high-voltage cable partial discharge pattern recognition system is provided, comprising:

[0032] an acquisition module, configured to acquire partial discharge data of different high-voltage cable insulation defect types, extract partial discharge transient pulses from the partial discharge data to construct characteristic parameters of the pulses, and normalize the obtained characteristic parameters to obtain original characteristic data;

[0033] A feature extraction module is used to extract features from the original feature data obtained in step 1) using SDAE, DBN and CNN to obtain deep learning high-order features;

[0034] A training module is used to establish a DESCA-CatBoost classifier model, and train the DESCA-CatBoost classifier model using the deep learning high-order features to obtain a trained DESCA-CatBoost classifier model;

[0035] The prediction module is used to identify the partial discharge pattern of high-voltage cables using the trained DESCA-CatBoost classifier.

[0036] The process of using SDAE, DBN and CNN to extract features from original feature data and obtain deep learning high-order features is as follows:

[0037] The original feature data is input into SDAE, DBN and CNN respectively, and the features of the last hidden layer of SDAE and DBN and the output features of the last Flatten layer of CNN are extracted to construct high-order features for deep learning.

[0038] In a third aspect of the present invention, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the high-voltage cable partial discharge pattern recognition method when executing the computer program.

[0039] In a fourth aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method for identifying partial discharge patterns of high-voltage cables are implemented.

[0040] The present invention has the following beneficial effects:

[0041] During specific operation, the high-voltage cable partial discharge pattern recognition method, system, device and storage medium described in the present invention utilize the excellent feature extraction capabilities of the deep learning models SDAE, DBN and CNN, adopt the DESCA-CatBoost classifier to use the high-order features extracted by the three deep learning models, and simultaneously utilize the differential evolution-sin-cosine hybrid algorithm DESCA to optimize the various hyperparameters of CatBoost, and thereby construct a classifier to improve the recognition accuracy of the classifier. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The accompanying drawings, which constitute part of the present invention, are intended to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are intended to explain the present invention and do not constitute an undue limitation of the present invention. In the accompanying drawings:

[0043] Figure 1 is a flow chart of the present invention;

[0044] Figure 2This is a flowchart of optimizing the hyperparameters of CatBoost using the differential evolution-sine-cosine hybrid algorithm DESCA in the present invention. DETAILED DESCRIPTION

[0045] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0046] It should be noted that the terms "first", "second", etc. in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0047] The present invention is described in further detail below with reference to the accompanying drawings:

[0048] Example 1

[0049] refer to Figure 1 The method for identifying partial discharge patterns of high-voltage cables according to the present invention comprises the following steps:

[0050] 1) Obtaining partial discharge data of different types of high-voltage cable insulation defects, extracting partial discharge instantaneous pulses from the partial discharge data to construct characteristic parameters of the pulses, and normalizing the characteristic parameters to obtain original characteristic data; the characteristic parameters of the pulses include discharge amount, peak voltage, average voltage, root mean square value, standard deviation, pulse width, rise time, fall time, equivalent time length, equivalent width, main frequency, wavelet characteristic parameters (ED1-ED3, EA3, Ea1-Ea3, Ed1-E3), phase angle, phase density, phase asymmetry, pulse polarity, skewness, kurtosis, crest factor, shape factor and partial discharge position. Among them, the wavelet feature parameter acquisition method is used to extract the partial discharge transient pulse. The specific process is: the signal is decomposed into three layers using wavelet transform, and the wavelet feature parameters of the signal are constructed according to the wavelet coefficients. ED1-ED3 are the detail energy ratio parameters of the three-layer signal after decomposition, Ea1-Ea3 are the global energy ratio parameters of the three-layer signal after decomposition, Ed1-Ed3 are the detail energy parameters of the three-layer signal after decomposition, and EA 3 is the global energy ratio parameter.

[0051] 2) Use SDAE, DBN and CNN to extract features from the original feature data obtained in step 1), extract the features of the last hidden layer of SDAE and DBN and the output features of the last Flatten layer of CNN, and use them to construct deep learning high-order features.

[0052] 3) establishing a DESCA-CatBoost classifier model, and using the deep learning high-order features obtained in step 2) as input to the DESCA-CatBoost classifier, training the DESCA-CatBoost classifier, and obtaining a trained DESCA-CatBoost classifier model;

[0053] CatBoost is a decision tree-based gradient boosting algorithm developed by Yandex researchers and engineers for various tasks such as search, recommendation systems, and weather forecasting. DE SCA-CatBoost uses the differential evolution-sine-cosine hybrid algorithm (DESCA) to optimize the hyperparameters of CatBoost. The specific process is as follows:

[0054] 31) Initialize the value range of each parameter of the CatBoost model and set the number of iterations of the DESCA overall optimization algorithm to T*;

[0055] 32) Randomly generate N populations, where the individuals in each population are equivalent to a set of CatBoost parameters (parameter 1, parameter 2, ..., parameter d);

[0056] 33) Using the classification accuracy as the individual fitness value, initialize the individual fitness values of all particles in the N populations in step 32);

[0057] 34) DE optimization is performed on each of these N populations to obtain N optimal particles. The DE optimization process is as follows: each population contains m individuals, the number of iterations for each population is set to T1, and mutation, crossover, and selection operations are performed on the m individuals to further update the population;

[0058] 35) Calculate the fitness value of each updated particle, and update the optimal individual representing the current number of iterations according to the size of the fitness value;

[0059] 36) Return to step 34) and continue to complete DE optimization until the upper limit of the number of iterations T1 is reached. Each population will have T1 historical optimal particles. The fitness of the T1 historical optimal particles is compared, and the particle with the highest fitness value is used as the optimal individual of the population. N optimal individuals from N populations are obtained;

[0060] 37) The N optimal individuals obtained in step 36) are used as the initial population of the SCA algorithm, and the improved SCA optimization is performed. The number of iterations of the SCA algorithm is set to T2;

[0061] 38) Initialize the parameters of the SCA algorithm, use the classification accuracy as the calculation formula for the fitness value, calculate the fitness value of the particle, retain the optimal particle Pg, and use formula (1), formula (2) and formula (3) to update the position of all particles:

[0062]

[0063] in, is the j-th dimension of the ith particle in the t-th iteration, is the optimal particle of the tth iteration, ω is the nonlinear weight; ω max is the maximum weight, the recommended value is 0.8, ω min is the minimum weight, the recommended value is 0.001, r1 is a linear decreasing function, r2∈[0,2π], r3∈[-2,2], r4∈[-1,1] are random numbers, a is a constant, generally 2, t is the current number of iterations, and T2 is the maximum number of iterations.

[0064] 39) Determine whether the current number of iterations is ≤ T2. If so, return to step 38); otherwise, jump to step 310);

[0065] 310) Determine whether the current total number of iterations is ≤ T*. If so, randomly select K historical optimal particles in SCA to replace the K individuals in each DE population in step 32), and return to step 32). Otherwise, output the optimal solution.

[0066] 4) The trained DESCA-CatBoost classifier is used to perform local discharge pattern recognition of high-voltage cables to achieve deep feature extraction of local discharge and defect type identification.

[0067] Confirmatory trials

[0068] Table 1 is a statistical table of the accuracy of the present invention and the prior art when classifying high-voltage cable partial discharge test sample data.

[0069] Table 1

[0070]

[0071]

[0072] Example 2

[0073] Accordingly, the present invention discloses a high-voltage cable partial discharge pattern recognition system, comprising:

[0074] an acquisition module, configured to acquire partial discharge data of different high-voltage cable insulation defect types, extract partial discharge transient pulses from the partial discharge data to construct characteristic parameters of the pulses, and normalize the obtained characteristic parameters to obtain original characteristic data;

[0075] A feature extraction module is used to extract features from the original feature data obtained in step 1) using SDAE, DBN and CNN to obtain deep learning high-order features;

[0076] A training module is used to establish a DESCA-CatBoost classifier model, and train the DESCA-CatBoost classifier model using the deep learning high-order features to obtain a trained DESCA-CatBoost classifier model;

[0077] The prediction module is used to identify the partial discharge pattern of high-voltage cables using the trained DESCA-CatBoost classifier.

[0078] The process of using SDAE, DBN and CNN to extract features from original feature data and obtain deep learning high-order features is as follows:

[0079] The original feature data is input into SDAE, DBN and CNN respectively, and the features of the last hidden layer of SDAE and DBN and the output features of the last Flatten layer of CNN are extracted to construct high-order features for deep learning.

[0080] Example 3

[0081] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the method for identifying partial discharge patterns in high-voltage cables are implemented. The memory may include internal memory, such as a high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device. The processor, network interface, and memory are interconnected via an internal bus, which may be an industrial standard architecture bus, a peripheral component interconnect standard bus, an extended industrial standard architecture bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. The memory is used to store programs. Specifically, the programs may include program code, which includes computer operating instructions. The memory may include internal memory and non-volatile memory, and provides instructions and data to the processor.

[0082] Example 4

[0083] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the method for identifying a partial discharge pattern in a high-voltage cable. Specifically, the computer-readable storage medium includes, but is not limited to, volatile memory and / or non-volatile memory. The volatile memory may include random access memory (RAM) and / or cache memory. The non-volatile memory may include read-only memory (ROM), a hard disk, flash memory, an optical disk, a magnetic disk, etc.

[0084] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0085] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1A device that provides the functions specified in a block or multiple blocks.

[0086] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0087] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0088] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the scope of protection of the claims of the present invention.

Claims

1. A method for identifying partial discharge patterns of high-voltage cables, characterized in that: The following steps are involved: Acquiring partial discharge data of different high-voltage cable insulation defect types, extracting partial discharge transient pulses from the partial discharge data to construct characteristic parameters of the pulses, and normalizing the characteristic parameters to obtain original characteristic data; Use SDAE, DBN and CNN to extract features from the original feature data to obtain deep learning high-order features; Establishing a DESCA-CatBoost classifier model, and training the DESCA-CatBoost classifier model using the deep learning high-order features to obtain a trained DESCA-CatBoost classifier model; The trained DESCA-CatBoost classifier is used to recognize partial discharge patterns of high-voltage cables.

2. The method for identifying partial discharge patterns of high-voltage cables according to claim 1, characterized in that: The characteristic parameters of the pulse include discharge amount, peak voltage, average voltage, root mean square value, standard deviation, pulse width, rise time, fall time, equivalent time length, equivalent width, main frequency, wavelet characteristic parameters, phase angle, phase density, phase asymmetry, pulse polarity, skewness, kurtosis, crest factor, shape factor and partial discharge position.

3. The method for identifying partial discharge patterns of high-voltage cables according to claim 1, characterized in that: The process of using SDAE, DBN and CNN to extract features from the original feature data obtained in step 1) to obtain deep learning high-order features is as follows: The original feature data is input into SDAE, DBN and CNN respectively, and the features of the last hidden layer of SDAE and DBN and the output features of the last Flatten layer of CNN are extracted to construct high-order features for deep learning.

4. The method for identifying partial discharge patterns of high-voltage cables according to claim 1, characterized in that: The process of constructing the DESCA-CatBoost classifier model is as follows: The differential evolution-sine-cosine hybrid algorithm DESCA is used to optimize the hyperparameters of CatBoost, and the optimized CatBoost is used to construct the DESCA-CatBoost classifier model.

5. The method for identifying partial discharge patterns of high-voltage cables according to claim 1, characterized in that: The process of optimizing the hyperparameters of CatBoost using the differential evolution-sine-cosine hybrid algorithm DESCA is as follows: 31) Initialize the value range of each parameter of the CatBoost model and set the number of iterations of the DESCA overall optimization algorithm to T*; 32) Randomly generate N populations, where the individuals in each population are equivalent to a set of CatBoost parameters; 33) Using the classification accuracy as the individual fitness value, initialize the individual fitness values of all particles in the N populations in step 32); 34) Performing a DE optimization on the N populations to obtain N optimal particles; 35) For the fitness values of the N optimal particles, update the optimal individual representing the current iteration number according to the size of the fitness value; 36) Return to step 34) and continue to complete DE optimization until the upper limit of the number of iterations T1 is reached. At this time, each population will have T1 historical optimal particles. The fitness of the T1 historical optimal particles is compared, and the particle with the highest fitness value is taken as the optimal individual of the population, and N optimal individuals from N populations are obtained; 37) The N best individuals from the N populations obtained in step 36) are used as the initial population of the SCA algorithm, and SCA optimization is performed. The number of iterations of the SCA algorithm is set to T2; 38) Initialize the parameters of the SCA algorithm, use the classification accuracy as the fitness value, calculate the fitness value of the particle, retain the optimal particle Pg, and use equations (1), (2) and (3) to update the positions of all particles: in, is the j-th dimension of the ith particle in the t-th iteration, is the optimal particle of the tth iteration, ω is the nonlinear weight, ω max is the maximum weight, ω min is the minimum weight, r1 is a linear decreasing function, r2∈[0,2π], r3∈[-2,2], r4∈[-1,1] are random numbers, a is a constant, t is the current number of iterations, and T2 is the maximum number of iterations; 39) Determine whether the current number of iterations is ≤ T2. If so, return to step 38); otherwise, jump to step 310); 310) Determine whether the current total number of iterations is ≤ T*. If so, randomly select K historical optimal particles in SCA to replace the K individuals in each DE population in step 32), and return to step 32). Otherwise, output the optimal solution.

6. A high-voltage cable partial discharge pattern recognition system, characterized in that: include: an acquisition module, configured to acquire partial discharge data of different high-voltage cable insulation defect types, extract partial discharge transient pulses from the partial discharge data to construct characteristic parameters of the pulses, and normalize the obtained characteristic parameters to obtain original characteristic data; A feature extraction module is used to extract features from the original feature data using SDAE, DBN and CNN to obtain deep learning high-order features; A training module is used to establish a DESCA-CatBoost classifier model, and train the DESCA-CatBoost classifier model using the deep learning high-order features to obtain a trained DESCA-CatBoost classifier model; The prediction module is used to identify the partial discharge pattern of high-voltage cables using the trained DESCA-CatBoost classifier.

7. The high-voltage cable partial discharge pattern recognition system according to claim 6, characterized in that: The process of using SDAE, DBN and CNN to extract features from original feature data and obtain deep learning high-order features is as follows: The original feature data is input into SDAE, DBN and CNN respectively, and the features of the last hidden layer of SDAE and DBN and the output features of the last Flatten layer of CNN are extracted to construct high-order features for deep learning.

8. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the method for identifying partial discharge patterns of high-voltage cables according to any one of claims 1 to 5 are implemented.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method for identifying partial discharge patterns of high-voltage cables according to any one of claims 1 to 5 are implemented.

Images