Data processing method and device, nonvolatile storage medium and computer device

Abstract

The application discloses a data processing method and device, a nonvolatile storage medium and computer equipment. The method comprises the following steps: measuring target partial discharge data of a target cable; inputting the target partial discharge data into a neural network model to output target space charge distribution data corresponding to the target partial discharge data, wherein the neural network model is a pre-trained full connection neural network model. The application solves the technical problem that the space charge distribution of a cable cannot be directly obtained from the partial discharge data of the cable in the prior art.

Description

Technical Field

[0001] This invention relates to the field of cable technology, and more specifically, to a data processing method, apparatus, non-volatile storage medium, and computer equipment. Background Technology

[0002] The assessment of high-voltage power equipment primarily focuses on the evaluation of its insulation. High-voltage power equipment experiences insulation degradation under prolonged exposure to high voltage and high temperature. This insulation degradation has a certain probability of affecting the operational efficiency of the power equipment and threatening the safe and stable operation of the power system.

[0003] Partial discharge and space charge distribution can indicate the insulation level of electrical equipment. Partial discharge develops rapidly and has a small time constant; space charge distribution changes slowly and has a large time constant. Considering that partial discharge may not necessarily occur in the early stages of insulation degradation, but changes in the distribution of internal electrical parameters (such as conductivity and dielectric constant) will lead to changes in the space charge distribution within the dielectric, space charge distribution, in principle, contains more information than partial discharge.

[0004] Space charge distribution measurement is difficult to perform offline, while partial discharge measurement can be performed online. There is no existing technology that can directly obtain the space charge distribution of a cable through partial discharge.

[0005] There is currently no effective solution to the above problems. Summary of the Invention

[0006] The present invention provides a data processing method, apparatus, non-volatile storage medium, and computer device to at least solve the technical problem that the spatial charge distribution of a cable cannot be obtained from the partial discharge data of the cable in the prior art.

[0007] According to one aspect of the present invention, a data processing method is provided, comprising: measuring target partial discharge data of a target cable; inputting the target partial discharge data into a neural network model, and outputting target space charge distribution data corresponding to the target partial discharge data, wherein the neural network model is a pre-trained fully connected neural network model.

[0008] Optionally, before inputting the target partial discharge data into the neural network model, the method further includes: acquiring a first partial discharge dataset and a first spatial charge distribution dataset, wherein the first partial discharge dataset and the first spatial charge distribution dataset are sample sets used to train the original neural network model, and one spatial charge data in the first spatial charge distribution dataset corresponds to multiple partial discharge data in the first partial discharge dataset; training the original neural network model using the first partial discharge dataset and the first spatial charge distribution dataset to obtain a neural network model, wherein the neural network model is used to output cable spatial distribution data corresponding to the cable partial discharge data based on the input cable partial discharge data.

[0009] Optionally, acquiring the first partial discharge dataset and the first space charge distribution dataset includes: acquiring multiple sets of correspondences between the space charge distribution data and partial discharge data of the experimental cable, wherein each set of correspondences includes one space charge distribution data corresponding to multiple partial discharge data; determining the multiple partial discharge data and multiple space charge data in the multiple sets of correspondences as the second partial discharge dataset and the second space charge distribution dataset, respectively; preprocessing the data in the second partial discharge dataset to obtain the third partial discharge dataset; inputting the third partial discharge dataset into the first encoder to obtain the first partial discharge dataset output by the first encoder, and inputting the second space charge distribution dataset into the second encoder to obtain the first space charge distribution dataset output by the second encoder, wherein the first encoder is used to extract the features of each partial discharge data in the third partial discharge dataset and generate a feature map of each partial discharge data, and the second encoder is used to extract the features of each space charge distribution data in the second space charge distribution dataset and generate a feature map of each space charge distribution data.

[0010] Optionally, the correspondence between multiple sets of space charge distribution data and partial discharge data of the experimental cable is obtained, including: measuring the space charge distribution data of the experimental cable and multiple partial discharge data corresponding to the space charge distribution data at different times, and obtaining the correspondence between multiple sets of space charge distribution data and partial discharge data of the experimental cable.

[0011] Optionally, the data in the second partial discharge dataset is preprocessed to obtain the third partial discharge dataset, including: determining the data type of each partial discharge data in the second partial discharge dataset; performing a Fourier transform on the partial discharge data of DC data type in the second partial discharge dataset to obtain the spectrum of the third partial discharge dataset; and plotting the phase spectrum of the partial discharge data of AC data type in the second partial discharge dataset to obtain the phase spectrum of the third partial discharge dataset.

[0012] Optionally, the target partial discharge data is input into a pre-trained neural network model, and the corresponding space charge distribution data is output. This includes: inputting the target partial discharge data into a first encoder to obtain a feature map of the target partial discharge data; inputting the feature map of the target partial discharge data into a pre-trained neural network model to output a feature map of the target space charge distribution data; and decoding the feature map of the target space charge distribution data through a first decoder to obtain the target space charge distribution data corresponding to the target partial discharge data. The first decoder and the second encoder belong to the same autoencoder.

[0013] Optionally, measuring the target partial discharge data of the target cable includes: measuring the original partial discharge data of the target cable; determining the data type of the original partial discharge data of the target cable; if the data type of the original partial discharge data is DC, performing a Fourier transform on the original partial discharge data to obtain the spectrum of the original partial discharge data; if the data type of the original partial discharge data is AC, plotting the phase spectrum of the original partial discharge data; and determining the spectrum or phase spectrum of the original partial discharge data as the target partial discharge data.

[0014] According to another aspect of the present invention, a data processing apparatus is also provided, comprising: a measurement module for measuring target partial discharge data of a target cable; and a processing module for inputting the target partial discharge data into a neural network model and outputting target space charge distribution data corresponding to the target partial discharge data, wherein the neural network model is a pre-trained fully connected neural network model.

[0015] According to another aspect of the present invention, a non-volatile storage medium is also provided, the non-volatile storage medium including a stored program, wherein, when the program is running, it controls the device where the non-volatile storage medium is located to execute any of the above data processing methods.

[0016] According to another aspect of the present invention, a computer device is also provided, the computer device including a processor, the processor being configured to run a program, wherein the program executes any of the above-described data processing methods during runtime.

[0017] In this embodiment of the invention, the target partial discharge data of the target cable is measured; the target partial discharge data is input into a neural network model, and the target space charge distribution data corresponding to the target partial discharge data is output. The neural network model is a pre-trained fully connected neural network model, which achieves the purpose of establishing a connection between the partial discharge data of the cable and the space charge distribution data of the cable using the neural network model. This achieves the technical effect of obtaining the corresponding space charge distribution data of the cable by inputting the partial discharge data of the cable into the neural network model, thereby solving the technical problem that the space charge distribution of the cable cannot be directly obtained from the partial discharge data of the cable in the prior art. Attached Figure Description

[0018] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0019] Figure 1 A hardware structure block diagram of a computer terminal for implementing a data processing method is shown.

[0020] Figure 2 This is a flowchart illustrating a data processing method provided according to an embodiment of the present invention;

[0021] Figure 3 This is a flowchart illustrating the data processing method provided by an optional embodiment of the present invention;

[0022] Figure 4 This is a schematic diagram of partial discharge data of a cable provided according to an optional embodiment of the present invention;

[0023] Figure 5 This is a schematic diagram of the space charge distribution data of a cable provided according to an optional embodiment of the present invention;

[0024] Figure 6 This is a schematic diagram of the space charge distribution data output by the neural network model provided in an optional embodiment of the present invention;

[0025] Figure 7 This is a structural block diagram of a data processing apparatus provided according to an embodiment of the present invention. Detailed Implementation

[0026] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0027] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0028] First, some nouns or terms that appear in the description of the embodiments of this application shall be interpreted as follows:

[0029] PyTorch is an open-source Python machine learning library used in applications such as natural language processing.

[0030] Tensorflow is a symbolic mathematics system based on dataflow programming, which is widely used in the programming implementation of various machine learning algorithms.

[0031] Stochastic Gradient Descent (SGD) is one of the earliest deep learning optimization methods. SGD and its derivatives use the same learning rate for optimization.

[0032] The Adaptive Momentum (Adam) stochastic optimization algorithm, based on the SGD algorithm, considers both first-order momentum and second-order momentum. The introduction of second-order momentum enables the adaptive learning rate.

[0033] According to an embodiment of the present invention, a method embodiment for data processing is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0034] The method embodiment provided in Embodiment 1 of this application can be executed on a mobile terminal, computer terminal, or similar computing device. Figure 1 A hardware structure block diagram of a computer terminal for implementing a data processing method is shown. Figure 1 As shown, the computer terminal 10 may include one or more processors (shown as 102a, 102b, ..., 102n in the figure) (the processor may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.) and a memory 104 for storing data. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of a BUS bus), a network interface, a power supply, and / or a camera. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the aforementioned electronic device. For example, computer terminal 10 may also include... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0035] It should be noted that the aforementioned one or more processors and / or other data processing circuits are generally referred to herein as "data processing circuits". These data processing circuits may be implemented wholly or partially as software, hardware, firmware, or any other combination thereof. Furthermore, the data processing circuits may be a single, independent processing module, or may be wholly or partially integrated into any other element in the computer terminal 10. As involved in the embodiments of this application, the data processing circuits serve as processor control (e.g., selection of a variable resistor termination path connected to an interface).

[0036] The memory 104 can be used to store software programs and modules of application software, such as program instructions / data storage devices corresponding to the data processing method in this embodiment of the invention. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory 104, thereby implementing the data processing method of the aforementioned application. The memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor, and these remote memories can be connected to the computer terminal 10 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0037] The display may be, for example, a touchscreen liquid crystal display (LCD) that allows the user to interact with the user interface of the computer terminal 10.

[0038] Figure 2 This is a flowchart illustrating a data processing method provided according to an embodiment of the present invention, such as... Figure 2 As shown, the method includes the following steps:

[0039] Step S202: Measure the target partial discharge data of the target cable.

[0040] In this step, the target cable is the one whose space charge distribution data needs to be determined using partial discharge data. The target cable can be a cable currently in use in the power grid. Directly measuring the space charge distribution data of the target cable is complex and inconvenient, but measuring its partial discharge data is relatively easier. However, when analyzing the operating condition of the target cable, the space charge distribution data contains more valuable information than the partial discharge data. Therefore, it is necessary to obtain the target space charge distribution data of the target cable by measuring its partial discharge data, which allows for a more accurate analysis of the target cable's operating condition.

[0041] Step S204: Input the target partial discharge data into the neural network model and output the target space charge distribution data corresponding to the target partial discharge data. The neural network model is a pre-trained fully connected neural network model.

[0042] In this step, after measuring the target partial discharge data of the target cable, the target partial discharge data is input into a pre-trained neural network model. The neural network model will output the target spatial charge distribution data corresponding to the target partial discharge data. The neural network model can be a fully connected neural network (MLP) model. The MLP model mainly consists of an input layer, a hidden layer, and an output layer. It is a neural network model constructed using a fully connected approach, meaning that every neuron in the previous layer is connected to every neuron in the next layer, and each layer is implemented using a fully connected approach.

[0043] Through the above steps, the goal of establishing a connection between the partial discharge data of the cable and the spatial charge distribution data of the cable using a neural network model can be achieved. This realizes the technical effect of obtaining the corresponding spatial charge distribution data of the cable by inputting the partial discharge data of the cable into the neural network model, thereby solving the technical problem that the spatial charge distribution of the cable cannot be directly obtained from the partial discharge data of the cable in the existing technology.

[0044] As an optional embodiment, before inputting the target partial discharge data into the neural network model, the following steps can be taken: obtaining a first partial discharge dataset and a first spatial charge distribution dataset, wherein the first partial discharge dataset and the first spatial charge distribution dataset are sample sets used to train the original neural network model, and one spatial charge data in the first spatial charge distribution dataset corresponds to multiple partial discharge data in the first partial discharge dataset; training the original neural network model using the first partial discharge dataset and the first spatial charge distribution dataset to obtain a neural network model, wherein the neural network model is used to output cable spatial distribution data corresponding to the cable partial discharge data based on the input cable partial discharge data.

[0045] Optionally, before inputting the target partial discharge data into the neural network model, the original neural network model needs to be trained first. This requires obtaining a sample set for training the original neural network model. The sample set consists of two parts: a first partial discharge dataset and a first spatial charge distribution dataset. The original neural network is trained using this sample set, enabling the trained model to output cable spatial distribution data corresponding to any input cable partial discharge data. In the sample set, one spatial charge distribution data point in the first spatial charge distribution dataset corresponds to multiple partial discharge data points in the first partial discharge dataset. The first spatial charge distribution dataset contains many spatial charge data points, each representing a spatial distribution of charge within the cable. Because the time constant of the spatial charge distribution data is large, while the time constant of the partial discharge data is small, one spatial charge distribution data point needs to correspond to multiple partial discharge data points.

[0046] The neural network model can be an MLP model, with an architecture structured as "input layer - several hidden layers - output layer". After training the MLP model, it can output cable spatial charge distribution data from the input partial discharge data of the cable. The input layer data is the first partial discharge dataset, with a data length of N*1. The output layer data is the first spatial charge distribution dataset, with a data length of M*1. The neural network model can be built and trained using PyTorch or TensorFlow frameworks. The loss function can be root mean square error, and the optimizer can be SGD or Adam. SGD and Adam are both commonly used optimizers for finding the global minimum. Adam converges faster and is more stable than SGD. However, for seeking a refined solution, SGD is recommended because although it takes longer to train and is prone to getting stuck in saddle points, with good initialization and learning rate scheduling, the results are more reliable.

[0047] As an optional embodiment, obtaining the first partial discharge dataset and the first space charge distribution dataset can be achieved through the following steps: obtaining multiple sets of correspondences between the space charge distribution data and partial discharge data of the experimental cable, wherein each set of correspondences includes one space charge distribution data corresponding to multiple partial discharge data; determining the multiple partial discharge data and multiple space charge data in the multiple sets of correspondences as the second partial discharge dataset and the second space charge distribution dataset, respectively; preprocessing the data in the second partial discharge dataset to obtain the third partial discharge dataset; inputting the third partial discharge dataset into the first encoder to obtain the first partial discharge dataset output by the first encoder, and inputting the second space charge distribution dataset into the second encoder to obtain the first space charge distribution dataset output by the second encoder, wherein the first encoder is used to extract the features of each partial discharge data in the third partial discharge dataset and generate a feature map of each partial discharge data, and the second encoder is used to extract the features of each space charge distribution data in the second space charge distribution dataset and generate a feature map of each space charge distribution data.

[0048] Optionally, to obtain the sample set for the neural network model, experiments can be conducted first to obtain multiple sets of correspondences between the space charge distribution data and partial discharge data of the experimental cable. Each set of correspondences corresponds to multiple partial discharge data in a space charge distribution data domain. The multiple partial discharge data obtained through multiple sets of correspondences are determined as the second partial discharge dataset. The data in the second partial discharge dataset needs to be preprocessed. The preprocessed second partial discharge dataset is the third partial discharge dataset. The third partial discharge dataset is input into the first encoder. The first encoder can extract the features of each partial discharge data in the third partial discharge dataset and generate a feature map, i.e., the first partial discharge dataset. The multiple space charge distribution data obtained through multiple sets of correspondences are determined as the second space charge distribution dataset, and the second space charge distribution dataset is input into the second encoder. The second encoder can extract the features of each space charge distribution dataset in the second space charge distribution dataset and generate a feature map, i.e., the first space charge distribution dataset. It should be noted that the partial discharge data and space charge distribution data of the experimental cable obtained in the experiment are not directly used as the sample set for training the original neural network model. Instead, they need to undergo some data processing. Specifically, the partial discharge data needs to be preprocessed before being input into the first encoder to extract the feature map before it can be used as the sample set, while the space charge distribution data needs to be input into the second encoder to extract the feature map before it can be used as the sample set.

[0049] As an optional embodiment, the correspondence between multiple sets of space charge distribution data and partial discharge data of the experimental cable can be obtained through the following steps: at different times, measure the space charge distribution data of the experimental cable and multiple partial discharge data corresponding to the space charge distribution data to obtain the correspondence between multiple sets of space charge distribution data and partial discharge data of the experimental cable.

[0050] Optionally, the experiment obtains the correspondence between multiple sets of space charge distribution data and partial discharge data. The space charge distribution data and partial discharge data of the experimental cable can be measured at different times, and the correspondence between the space charge distribution data and partial discharge data can be recorded. During the experiment, the experimental cable can be placed in an aging oven to accelerate insulation degradation. The experimental cable is taken out for testing at fixed intervals, and each test yields one space charge distribution data point and multiple partial discharge data points. Alternatively, the space charge distribution data of the experimental cable can be measured multiple times for each test, and the average value is taken as the final space charge distribution data for that test.

[0051] As an optional embodiment, the data in the second partial discharge dataset is preprocessed to obtain the third partial discharge dataset. This can be achieved through the following steps: determining the data type of each partial discharge data in the second partial discharge dataset; performing a Fourier transform on the partial discharge data in the second partial discharge dataset that is of DC data type to obtain the spectrum of the third partial discharge dataset; and plotting the phase spectrum of the partial discharge data in the second partial discharge dataset that is of AC data type to obtain the phase spectrum of the third partial discharge dataset.

[0052] Optionally, preprocessing the data in the second partial discharge dataset involves denoising the collected partial discharge data and space charge distribution data. The characteristics of the collected partial discharge data are complex. The most obvious characteristic is that when no partial discharge occurs, the device samples background white noise. This white noise data does not belong to the category of partial discharge characteristics. If the collected partial discharge data is directly used as the input of the first encoder, it will lead to a large deviation in the model. Therefore, preprocessing of the collected partial discharge data is necessary. Before preprocessing the partial discharge data, it is necessary to confirm the type of partial discharge data, i.e., whether the partial discharge data was measured in a DC system or an AC system. When the partial discharge data is collected in an AC system, the data in the collected second partial discharge dataset needs to be processed to obtain a phase spectrum (PRPD spectrum). The phase spectrum can reflect most of the partial discharge information from levels such as the maximum discharge amount within the period, the average discharge amount within the period, and the number of discharges within the period. When partial discharge data is acquired in a DC system, since there is no AC excitation voltage, the phase information cannot be used to calibrate the partial discharge data. In this case, the time-domain partial discharge data is converted to the frequency domain through Fourier transform to obtain the spectrum of the partial discharge data.

[0053] As an optional embodiment, inputting the target partial discharge data into a pre-trained neural network model and outputting the space charge distribution data corresponding to the target partial discharge data can be achieved through the following steps: inputting the target partial discharge data into a first encoder to obtain the feature map of the target partial discharge data; inputting the feature map of the target partial discharge data into a neural network model to output the feature map of the target space charge distribution data; decoding the feature map of the target space charge distribution data through a first decoder to obtain the target space charge distribution data corresponding to the target partial discharge data, wherein the first decoder and the second encoder belong to the same autoencoder.

[0054] Optionally, after measuring the target partial discharge data of the target cable, the target partial discharge data needs to be input into the first encoder. The first encoder extracts the feature map of the target partial discharge data, which is then input into the trained neural network model. The neural network model outputs a feature map of the target spatial charge distribution data corresponding to the feature map of the target partial discharge data. The first decoder decodes the feature map of the target spatial charge distribution data into the target spatial charge distribution data. The first decoder and the second encoder belong to the same autoencoder. When training the original neural network model to establish the sample set, the second encoder in this autoencoder extracts the feature map of the measured spatial charge distribution data. After the neural network model outputs the feature map of the target spatial charge distribution data, the first decoder of this autoencoder can decode the feature map to obtain the target spatial charge distribution data.

[0055] As an optional embodiment, measuring the target partial discharge data of the target cable can be achieved through the following steps: measuring the original partial discharge data of the target cable; determining the data type of the original partial discharge data of the target cable; if the data type of the original partial discharge data is DC, performing a Fourier transform on the original partial discharge data to obtain the spectrum of the original partial discharge data; if the data type of the original partial discharge data is AC, plotting the phase spectrum of the original partial discharge data; and determining the spectrum or phase spectrum of the original partial discharge data as the target partial discharge data.

[0056] Optionally, the target partial discharge data for the target cable can be pre-processed partial discharge data. When measuring the target cable, the original partial discharge data of the target cable is collected, allowing determination of the type of original partial discharge data—whether it was measured in a DC system or an AC system. When the original partial discharge data is collected in an AC system, it needs to be processed to obtain a PRPD spectrum. When the original partial discharge data is collected in a DC system, since there is no AC excitation voltage, phase information cannot be used to calibrate the original partial discharge data. In this case, the time-domain partial discharge data is transformed to the frequency domain using a Fourier transform to obtain the spectrum of the original partial discharge data. The final phase spectrum or spectrum can then be used as the target partial discharge data and input into an autoencoder for feature extraction.

[0057] As a specific embodiment, Figure 3 This is a flowchart illustrating the implementation of a data processing method according to an optional embodiment of the present invention, such as... Figure 3As shown in the figure, the present invention will realize the association of the partial discharge database and the space charge database. By only measuring the partial discharge data of the cable, the space charge distribution of the cable can be deduced to a certain extent. Since the partial discharge data and the space charge distribution are mostly used to predict the deterioration of electrical insulation, when selecting samples, it is only necessary to design an experimental cable with a fixed defect type; since the deterioration of electrical insulation has a long time span, it is necessary to combine aging tests to accelerate the process of data acquisition.

[0058] The specific implementation steps are as Figure 3 shown. First, it is necessary to determine the type of the research defect. After determining the defect type, record the relevant data of the measured defect (type, defect size, etc.); place the defect in an aging oven for aging to accelerate the process of insulation deterioration. Every fixed time interval (here, 12 hours is selected), place the defective sample on the combined partial discharge and space charge testing device for testing. Each time, collect multiple partial discharge data and one space charge distribution data; after a large number of experiments, obtain the second partial discharge data set and the second space charge distribution data set. It should be noted that when forming the second partial discharge data set and the second space charge distribution data set, if the partial discharge data is not measured, this set of partial discharge data and space charge data will not be incorporated into the data set.

[0059] Secondly, use the collected second partial discharge data set and the second space charge data set for model training. The model training process is as follows: If the test system is an AC system, process the second partial discharge data into a PRPD pattern to obtain the third partial discharge pattern data set; if the test system is a DC system, process the original partial discharge data set through Fourier transform to obtain the third partial discharge pattern data set.

[0060] Taking the partial discharge spectrum or the partial discharge phase pattern as the input, adopt unsupervised learning means to train the autoencoder AE1: The input of the autoencoder AE1 is the partial discharge spectrum or the phase pattern. Assume that the length of the spectrum data of a single input is L1*1, and the output end of the encoder in the autoencoder is designed as N1*1 (N < L1). Then, through the encoder of the trained autoencoder AE1, the partial discharge signal can be extracted as N1-dimensional partial discharge feature data, that is, Figure 3 the partial discharge abstract data set described in. Taking the space charge distribution data as the input, adopt unsupervised learning means to train the autoencoder AE2: The input of the autoencoder AE2 is the space charge distribution data. Assume that the length of the single input space charge distribution data is L2*L3, and the output end of the encoder in the autoencoder is designed as N2*1 (N2 < L2*L3). Then, through the encoder of the trained autoencoder AE2, the space charge distribution data can be extracted as N2-dimensional space charge distribution feature data, that is, Figure 3 the space charge abstract data set described in.

[0061] Secondly, using the first partial discharge dataset as input and the first space charge dataset as output, an MLP neural network from N1-dimensional to N2-dimensional is trained to obtain the MLP model, realizing the mapping from the N1-dimensional partial discharge dataset to the N2-dimensional space charge dataset.

[0062] Finally, after the model training is completed, it can be put into use. The specific method is as follows: the collected raw partial discharge data is first preprocessed to obtain the target partial discharge data, then passed through the encoder of the trained autoencoder AE1 to obtain the feature map of the target partial discharge data, the feature map of the target partial discharge data is input into the MLP model to obtain the feature map of the target space charge distribution data, and the feature map of the target space charge distribution data is passed through the decoder of the autoencoder AE2 to obtain the space charge distribution corresponding to the partial discharge signal (e.g., ...). Figure 3 (As shown by the dashed line).

[0063] Figure 4 This is a schematic diagram of partial discharge data of a cable provided according to an optional embodiment of the present invention, such as... Figure 4 The image shows the original partial discharge data of the target cable measured in a specific embodiment. Figure 5 This is a schematic diagram of the space charge distribution data of a cable provided according to an optional embodiment of the present invention, such as... Figure 5 The space charge distribution data shown are Figure 4 The actual space charge distribution data corresponding to the partial discharge data. Figure 6 This is a schematic diagram of the space charge distribution data output by the neural network model provided in an optional embodiment of the present invention, such as... Figure 6 The space charge distribution shown is Figure 4 The partial discharge data in the middle is processed through any of the above methods and then used to obtain the space charge distribution map of the target cable via a neural network model. By comparison... Figure 5 and Figure 6 As can be seen, the distribution of space charge shows relatively small differences, proving that the data processing method provided by this invention is feasible.

[0064] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the present invention is not limited to the described order of actions, because according to the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to the present invention.

[0065] Through the above description of the embodiments, those skilled in the art can clearly understand that the data processing method according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platform. Of course, it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0066] According to embodiments of the present invention, an apparatus for implementing the above-described data processing method is also provided. Figure 7 This is a structural block diagram of a data processing apparatus provided according to an embodiment of the present invention, such as... Figure 7 As shown, the data processing device includes a measurement module 72 and a processing module 74. The device will be described below.

[0067] Measurement module 72 is used to measure the target partial discharge data of the target cable.

[0068] The processing module 74, connected to the measurement module 72, is used to input the target partial discharge data into the neural network model and output the target space charge distribution data corresponding to the target partial discharge data. The neural network model is a pre-trained fully connected neural network model.

[0069] It should be noted that the measurement module 72 and processing module 74 mentioned above correspond to steps S202 to S204 in the embodiments. The two modules and their corresponding steps implement the same instances and application scenarios, but are not limited to the content disclosed in the above embodiments. It should also be noted that the above modules, as part of the device, can run in the computer terminal 10 provided in the embodiments.

[0070] Embodiments of the present invention may provide a computer device. Optionally, in this embodiment, the computer device may be located in at least one of a plurality of network devices in a computer network. The computer device includes a memory and a processor.

[0071] The memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the data processing method and apparatus in this embodiment of the invention. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, thereby realizing the aforementioned data processing method. The memory may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include memory remotely located relative to the processor, and these remote memories can be connected to a computer terminal via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0072] The processor can call the information and application program stored in the memory through the transmission device to perform the following steps: measuring the target partial discharge data of the target cable; inputting the target partial discharge data into the neural network model and outputting the target space charge distribution data corresponding to the target partial discharge data, wherein the neural network model is a pre-trained fully connected neural network model.

[0073] This invention provides a data processing solution. It measures the target partial discharge data of a target cable; inputs the target partial discharge data into a neural network model; and outputs the target space charge distribution data corresponding to the target partial discharge data. The neural network model is a pre-trained fully connected neural network model. This achieves the goal of establishing a connection between the cable's partial discharge data and its space charge distribution data using the neural network model. This realizes the technical effect of obtaining the corresponding space charge distribution data of the cable by inputting the cable's partial discharge data into the neural network model, thereby solving the technical problem that the space charge distribution of a cable cannot be obtained from its partial discharge data in existing technologies.

[0074] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing the hardware related to the terminal device. The program can be stored in a non-volatile storage medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk, etc.

[0075] Embodiments of the present invention also provide a non-volatile storage medium. Optionally, in this embodiment, the non-volatile storage medium can be used to store the program code executed by the data processing method provided in the above embodiments.

[0076] Optionally, in this embodiment, the non-volatile storage medium may be located in any computer terminal in a group of computer terminals in a computer network, or in any mobile terminal in a group of mobile terminals.

[0077] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: measuring target partial discharge data of the target cable; inputting the target partial discharge data into a neural network model and outputting target space charge distribution data corresponding to the target partial discharge data, wherein the neural network model is a pre-trained fully connected neural network model.

[0078] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0079] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0080] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0081] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0082] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0083] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a non-volatile storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0084] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A data processing method, characterized in that, Including: Measuring the target partial discharge data of the target cable; Inputting the target partial discharge data into a neural network model, and outputting the target space charge distribution data corresponding to the target partial discharge data, wherein the neural network model is a pre-trained fully connected neural network model; Wherein, before inputting the target partial discharge data into the neural network model, it further includes: obtaining a first partial discharge data set and a first space charge distribution data set, wherein the first partial discharge data set and the first space charge distribution data set are sample sets for training the original neural network model, and one space charge data in the first space charge distribution data set corresponds to multiple partial discharge data in the first partial discharge data set; training the original neural network model with the first partial discharge data set and the first space charge distribution data set to obtain the neural network model, wherein the neural network model is used to output the cable space distribution data corresponding to the input cable partial discharge data; wherein, the first partial discharge data set is obtained by preprocessing the partial discharge data of the experimental cable measured in the experiment and then inputting it into the first encoder to extract the feature map, and the first space charge distribution data set is obtained by inputting the space charge distribution data of the experimental cable measured in the experiment into the second encoder to extract the feature map; Wherein, the training process of the first encoder and the second encoder is: performing Fourier transform on the partial discharge data to obtain the partial discharge spectrum, using the partial discharge spectrum as the input, and training the autoencoder AE1 by means of unsupervised learning, including: inputting the partial discharge spectrum into the autoencoder AE1, the length of the input spectrum data is L1*1, the output end of the encoder of the autoencoder AE1 is designed as N1*1, wherein N1 < L1, and using the encoder of the trained autoencoder AE1 as the first encoder, which is used to extract the partial discharge data into N1-dimensional partial discharge feature data as the first partial discharge data set; using the space charge distribution data as the input, and training the autoencoder AE2 by means of unsupervised learning, including: inputting the space charge distribution data into the autoencoder AE2, the length of the input space charge distribution data is L2*L3, the output end of the encoder of the autoencoder AE2 is designed as N2*1, wherein N2 < L2*L3, and using the encoder of the trained autoencoder AE2 as the second encoder, which is used to extract the space charge distribution data into N2-dimensional space charge distribution feature data as the first space charge distribution data set.

2. The method according to claim 1, characterized in that, The obtaining of the first partial discharge data set and the first space charge distribution data set includes: Obtaining multiple groups of corresponding relationships between the space charge distribution data and the partial discharge data of the experimental cable, wherein each corresponding relationship in the multiple groups of corresponding relationships includes one space charge distribution data corresponding to multiple partial discharge data; The multiple partial discharge data and multiple space charge data in the multiple sets of correspondences are respectively determined as the second partial discharge dataset and the second space charge distribution dataset; The data in the second partial discharge dataset is preprocessed to obtain the third partial discharge dataset; The third partial discharge dataset is input into the first encoder to obtain the first partial discharge dataset output by the first encoder, and the second space charge distribution dataset is input into the second encoder to obtain the first space charge distribution dataset output by the second encoder. The first encoder is used to extract the features of each partial discharge data in the third partial discharge dataset and generate a feature map of each partial discharge data. The second encoder is used to extract the features of each space charge distribution data in the second space charge distribution dataset and generate a feature map of each space charge distribution data.

3. The method according to claim 2, characterized in that, The process of obtaining the correspondence between multiple sets of space charge distribution data and partial discharge data of the experimental cable includes: At different times, the space charge distribution data of the experimental cable and multiple partial discharge data corresponding to the space charge distribution data are measured to obtain the correspondence between multiple sets of space charge distribution data and partial discharge data of the experimental cable.

4. The method according to claim 2, characterized in that, The step of preprocessing the data in the second partial discharge dataset to obtain the third partial discharge dataset includes: Determine the data type of each partial discharge data in the second partial discharge dataset; Perform a Fourier transform on the DC-type partial discharge data in the second partial discharge dataset to obtain the spectrum of the third partial discharge dataset. Phase maps of AC partial discharge data from the second partial discharge dataset are plotted to obtain the phase maps of the third partial discharge dataset.

5. The method according to claim 2, characterized in that, The step of inputting the target partial discharge data into a pre-trained neural network model and outputting the space charge distribution data corresponding to the target partial discharge data includes: The target partial discharge data is input into the first encoder to obtain the feature map of the target partial discharge data; The feature map of the target partial discharge data is input into the neural network model, and the feature map of the target space charge distribution data is output. The feature map of the target space charge distribution data is decoded by the first decoder to obtain the target space charge distribution data corresponding to the target partial discharge data, wherein the first decoder and the second encoder belong to the same autoencoder.

6. The method according to claim 1, characterized in that, The target partial discharge data of the measured target cable includes: Measure the raw partial discharge data of the target cable; Determine the data type of the raw partial discharge data of the target cable; If the data type of the original partial discharge data is DC, perform a Fourier transform on the original partial discharge data to obtain the spectrum of the original partial discharge data; If the data type of the original partial discharge data is AC, draw the phase spectrum of the original partial discharge data; The spectrum of the original partial discharge data or the phase spectrum of the original partial discharge data is determined as the target partial discharge data.

7. A data processing apparatus, characterized in that, include: A measurement module for measuring target partial discharge data of a target cable; A processing module for inputting the target partial discharge data into a neural network model and outputting target space charge distribution data corresponding to the target partial discharge data, where the neural network model is a pre-trained fully connected neural network model; Wherein, the device is further configured to, before inputting the target partial discharge data into the neural network model, obtain a first partial discharge data set and a first space charge distribution data set, where the first partial discharge data set and the first space charge distribution data set are sample sets for training an original neural network model, and one space charge data in the first space charge distribution data set corresponds to multiple partial discharge data in the first partial discharge data set; use the first partial discharge data set and the first space charge distribution data set to train the original neural network model to obtain the neural network model, where the neural network model is used to output cable space distribution data corresponding to the input cable partial discharge data; wherein, the first partial discharge data set is obtained by preprocessing the partial discharge data of an experimental cable measured in an experiment and then inputting it into a first encoder to extract a feature map, and the first space charge distribution data set is obtained by inputting the space charge distribution data of the experimental cable measured in the experiment into a second encoder to extract a feature map; Wherein, the training processes of the first encoder and the second encoder are as follows: performing Fourier transform on the partial discharge data to obtain a partial discharge spectrum, using the partial discharge spectrum as an input, and training an autoencoder AE1 by means of unsupervised learning, including: inputting the partial discharge spectrum into the autoencoder AE1, the length of the input spectrum data being L1*1, the output end of the encoder of the autoencoder AE1 being designed as N1*1, where N1 < L1, and using the encoder of the trained autoencoder AE1 as the first encoder for extracting the partial discharge data into N1-dimensional partial discharge feature data as the first partial discharge data set; using the space charge distribution data as an input and training an autoencoder AE2 by means of unsupervised learning, including: inputting the space charge distribution data into the autoencoder AE2, the length of the input space charge distribution data being L2*L3, the output end of the encoder of the autoencoder AE2 being designed as N2*1, where N2 < L2*L3, and using the encoder of the trained autoencoder AE2 as the second encoder for extracting the space charge distribution data into N2-dimensional space charge distribution feature data as the first space charge distribution data set.

8. A non-volatile storage medium, characterized in that, The non-volatile storage medium includes a stored program, wherein when the program runs, it controls the device where the non-volatile storage medium is located to execute the data processing method according to any one of claims 1 to 6.

9. A computer device, characterized in that, Including: A memory and a processor, The memory stores a computer program; The processor is configured to execute a computer program stored in the memory, wherein when the computer program is executed, the processor performs the data processing method according to any one of claims 1 to 6.

Images