CN116027158A

CN116027158A - High-voltage cable partial discharge fault prediction method and system

Info

Publication number: CN116027158A
Application number: CN202310076323.8A
Authority: CN
Inventors: 刘佳鑫; 王帅; 周榆晓; 张国钢; 郑伟; 赵陈琛; 赵子健; 鲁旭臣; 崔巨勇; 王雅楠
Original assignee: State Grid Corp of China SGCC; Xian Jiaotong University; Electric Power Research Institute of State Grid Liaoning Electric Power Co Ltd
Current assignee: State Grid Corp of China SGCC; Xian Jiaotong University; Electric Power Research Institute of State Grid Liaoning Electric Power Co Ltd
Priority date: 2023-02-07
Filing date: 2023-02-07
Publication date: 2023-04-28

Abstract

The invention discloses a high-voltage cable partial discharge fault prediction method and a system, which are used for collecting multiple items of partial discharge data, temperature data and power data of on-line monitoring of a cable and realizing the positioning of the on-line monitoring of the partial discharge of the high-voltage cable; extracting morphological statistical feature quantity in an online acquisition PRPD map; filling the missing items of the rest monitoring parameters of the time sequence missing; respectively extracting each trend item and each period item by using the rest monitoring parameters and the extracted partial discharge morphological statistical characteristic quantity, and then combining and constructing a trend prediction model; and taking the trend prediction results of the rest monitoring parameters and the partial discharge form statistics characteristic quantity as the input of a partial discharge diagnosis model to obtain a high-voltage cable partial discharge fault diagnosis result. The prediction, positioning and diagnosis of the high-voltage cable partial discharge faults are realized. The invention realizes real-time positioning and auxiliary decision making of the insulation operation risk of the high-voltage cable, and is beneficial to improving the reliability early warning, global perception, strong robust regulation and control and other multi-aspect capabilities of the digital twin of the power system.

Description

High-voltage cable partial discharge fault prediction method and system

Technical Field

The invention belongs to the technical field of power equipment state prediction and fault diagnosis, and particularly relates to a high-voltage cable partial discharge fault prediction method and system.

Background

In order to adapt to the demand trend of clean energy development and low-carbon transformation, a large number of new energy power plants are connected to the power grid. Grid safety, stability and operation optimization will be an important challenge for the approach, especially for "dual high" power systems, which presents unprecedented challenges for conventional prediction of grid operating conditions and transmission control. And with the improvement of the requirements of social construction on the operation safety and reliability of the power system, the multi-dimensional detection, state prediction and system management of the power system are perfected so as to become the main attack direction and demand of the current power grid digitization.

For data, along with the construction and development of the electric power Internet of things, massive historical data are accumulated in the construction and operation processes of power grid equipment. The accumulated precipitation of a large amount of equipment data is in a 'deep sleep state', the value of the equipment data cannot be effectively mined, a set of means capable of carrying out scientific analysis and efficient decision on numerous data such as historical tests and online detection is lacked, and the existing production operation is difficult to be accurately guided by utilizing the large data. The high-voltage cable is used as important transmission electric equipment, and relates to a large number of relevant parameters representing the operation state of the high-voltage cable, such as on-line monitoring data, electric test data, power grid operation data, meteorological environment data, insulation monitoring data, equipment quality records and the like. And because of mutual coupling and close relation among the characteristic attributes. The cable insulation problem is always a technical core of stable and reliable operation of the cable, so that the current state of the high-voltage cable is diagnosed and positioned in an omnibearing manner through a data driving method, and the future development trend of the state is predicted and diagnosed, thereby being beneficial to rapidly, effectively and pertinently arranging equipment maintenance and formulating operation and maintenance strategies.

Disclosure of Invention

The invention aims to solve the technical problems of incapability of predicting the partial discharge fault state of a high-voltage cable, diagnosis of the type of the partial discharge fault and positioning of the type of the partial discharge fault.

The invention adopts the following technical scheme:

a high-voltage cable partial discharge fault prediction method comprises the following steps:

s1, collecting state data of a high-voltage cable, and constructing a cable partial discharge fault positioning data set, a cable state prediction historical data set and an original cable fault diagnosis historical data set;

s2, carrying out inverse reconstruction on the high-frequency signals in the cable partial discharge fault location data set obtained in the step S1 by utilizing wavelet filtering to obtain high-frequency signals;

s3, extracting modal components of the high-frequency signal obtained in the step S2, and obtaining a primary limited bandwidth inherent modal function of the high-frequency signal;

s4, based on a double-end phase response method, processing the high-frequency signals in the cable partial discharge fault location data set in the step S1 by utilizing fast Fourier distribution change to obtain a corresponding phase response map, and simultaneously, realizing the phase response of the peak value in the high-frequency signal primary limited bandwidth inherent mode function obtained in the step S3, and realizing the fault location of the high-voltage cable partial discharge by utilizing a cable partial discharge phase attenuation characteristic equation;

S5, extracting morphological statistical feature quantity of the PRPD map in the cable state prediction historical data set obtained in the step S1, and replacing the PRPD map by using the morphological statistical feature quantity of the PRPD map to form a new cable state prediction historical data set;

s6, filling the data of each cable characteristic parameter in the new cable state prediction historical data set obtained in the step S5 by utilizing multiple linear regression to form a cable state historical data set;

s7, decomposing each item of historical data in the cable state historical data set obtained in the step S6 into a trend component, a period component and a remainder by using a time sequence trend decomposition algorithm, and then realizing the combined time sequence prediction of each item of historical data trend component by using a long-short-term memory network to obtain a prediction result of each item of characteristic parameters of the high-voltage cable;

s8, extracting morphological statistical feature quantity of the PRPD map in the cable fault diagnosis historical data set obtained in the step S1, and replacing the PRPD map by using the morphological statistical feature quantity of the PRPD map to form a new cable fault diagnosis historical data set;

s9, establishing a fault diagnosis model of the deep convolution confidence network by using the new cable fault diagnosis historical data set obtained in the step S8, storing a model with an optimal training result as a DCBN cable fault diagnosis model, and taking the prediction result of each characteristic parameter of the high-voltage cable obtained in the step S7 as the input of the DCBN cable fault diagnosis model to realize the prediction of the recent partial discharge fault of the cable.

Specifically, in step S1, PRPD spectrum, high-frequency signal, ground current, temperature and load current data of the high-voltage cable are collected; taking high-frequency signals acquired by two high-frequency sensors adjacent to each other on a high-voltage cable as a cable partial discharge fault positioning data set; taking the PRPD map, the grounding current, the temperature and the load current as a cable state prediction historical data set; and collecting the data of the PRPD map, the grounding current, the temperature and the load current of the high-voltage cable in normal state and the generation of partial drop, and forming an original cable fault diagnosis historical data set.

Specifically, in step S3, a Lagrange multiplier method and a penalty factor α are introduced to make the partial discharge high-frequency signal variational decomposition have constraint, and a Lagrange multiplier factor and an update factor σ are used to iteratively optimize through an alternate direction algorithm until the convergence criterion fault tolerance epsilon is satisfied, so as to obtain a series of finite bandwidth intrinsic mode functions after respective decomposition, and finally obtain the primary finite bandwidth intrinsic mode functions of the two high-frequency signals.

Specifically, step S4 specifically includes:

acquiring two high-frequency signals adjacent to each other on a high-voltage cable; performing fast Fourier transform on the two high-frequency signals to obtain phase response maps of the two high-frequency signals; respectively acquiring phase response of peak value in principal finite bandwidth natural mode functions of two high-frequency signals

And->

Simultaneous->

And->

And determining the position of the partial discharge fault.

Specifically, in step S5, PRPD profile feature extraction includes a bias S _k Steepness K _u A discharge factor Q, a correlation coefficient CC, and a phase asymmetry ψ.

Specifically, in step S6, the data missing patch of the multiple linear regression is specifically:

s601, acquiring a complete parameter set in historical data, arranging and combining different parameters, fitting all combined relations by using multiple linear regression, acquiring data acquired by an online sensor, and determining the type of data missing at a certain moment;

s602, using missing data as a dependent variable ζ, using non-missing data as an independent variable alpha, using historical data to approach alpha beta plus epsilon=ζ, wherein beta is a weight matrix, and epsilon is a bias matrix;

s603, filling the missing items by using the fitting result.

Specifically, in step S7, the joint timing prediction is specifically:

s701, acquiring historical time sequence data of a plurality of parameters, and performing STL decomposition on the map statistical characteristic quantity, the grounding current, the temperature and the load current respectively to obtain trend components of the historical data of the characteristic parameters;

s702, definitely needing predicted time span;

s703, taking the time span of the period quantity obtained by STL decomposition as the length of an input window of LSTM;

S704, utilizing an LSTM model, and combining trend components obtained by decomposing all parameters in the step S701 to serve as input of combined time sequence prediction, so as to obtain a trend component time sequence prediction result of each characteristic parameter of the cable;

and S705, adding the trend prediction results of the different parameters obtained in the step S704 to the respective period components to realize multi-parameter joint timing prediction.

Specifically, in step S9, the fault diagnosis using the deep convolutional confidence network DCBN is specifically:

normalizing the cable fault diagnosis historical data set obtained in the step S8 to be [ -1,1] as input of a DCBN cable fault diagnosis model, and performing one-hot encoding on the partial discharge type as output;

setting interface parameters of a DCBN cable fault diagnosis model, namely the number of neurons of an input layer and the number of neurons of an output layer, wherein the number of the input layer is equal to the number of characteristic parameters, and the number of the output layer is equal to the number of partial discharge types;

setting an internal network structure of a DCBN cable fault diagnosis model, extracting an output effective characteristic value by utilizing CNN at the front end of the internal network, and performing unsupervised optimization on a plurality of layers of RBM of the internal network structure; the method comprises the steps of performing global optimization on an internal network structure by using a BP algorithm by using full connection of the tail end of the internal network structure;

Normalizing the multi-parameter joint time sequence prediction result obtained in the step S7 by using the normalized scale of the historical dataset; and inputting the multi-parameter joint time sequence prediction result into an optimal DCBN cable fault diagnosis model to obtain a final diagnosis result and realize cable partial discharge fault prediction.

Further, the optimal DCBN cable fault diagnosis model is specifically:

acquiring a historical fault data set, and determining the number of neurons of an input layer and an output layer, the number of layers of RBM (radial basis function) of a network structure, the number of neurons of each layer of RBM, parameters of a convolution layer and a pooling layer, and convolution kernel and padding; and then performing one-hot coding on the partial discharge fault type of the data set to be used as output of a DCBN cable fault diagnosis model, normalizing characteristic parameters to [ -1,1] and dividing a training set and a test set, training by using the training set and tuning by using the test set, and storing a model with the optimal effect of the test set as the DCBN cable fault diagnosis model.

In a second aspect, an embodiment of the present invention provides a high-voltage cable partial discharge fault prediction system, including:

the data module is used for collecting state data of the high-voltage cable, constructing a cable partial discharge fault positioning data set, a cable state prediction historical data set and an original cable fault diagnosis historical data set;

The reconstruction module is used for reversely reconstructing the high-frequency signals in the cable partial discharge fault positioning data set obtained by the wavelet filtering processing data module to obtain filtered high-frequency signals;

the extraction module is used for extracting modal components of the high-frequency signal obtained by the reconstruction module and obtaining a primary limited bandwidth inherent modal function of the high-frequency signal;

the positioning module is used for processing high-frequency signals in the cable partial discharge fault positioning data set in the data module by utilizing the fast Fourier distribution change based on the double-end phase response method to obtain a corresponding phase response map, and the phase response of the peak value in the high-frequency signal primary limited bandwidth inherent mode function obtained by the simultaneous extraction module is utilized to realize the fault positioning of the high-voltage cable partial discharge by utilizing the cable partial discharge phase attenuation characteristic equation;

the first replacing module extracts the morphological statistical feature quantity of the PRPD map in the cable state prediction historical data set obtained by the data module, and replaces the PRPD map by the morphological statistical feature quantity of the PRPD map to form a new cable state prediction historical data set;

the filling module is used for filling the data of each cable characteristic parameter in the new cable state prediction historical data set obtained by the first replacement module by utilizing multiple linear regression to form a cable state historical data set;

The decomposition module is used for decomposing each item of historical data in the cable state historical data set obtained by the filling module into a trend component, a period component and a remainder by using a time sequence trend decomposition algorithm, and then realizing the combined time sequence prediction of each item of historical data trend component by using a long-short-period memory network to obtain a prediction result of each item of characteristic parameters of the high-voltage cable;

the second replacing module extracts the morphological statistical characteristic quantity of the PRPD map in the cable fault diagnosis historical data set obtained by the data module, and replaces the PRPD map by the morphological statistical characteristic quantity of the PRPD map to form a new cable fault diagnosis historical data set;

and the prediction module establishes a fault diagnosis model of the deep convolution confidence network by utilizing the new cable fault diagnosis historical data set obtained by the second replacement module, saves a model with an optimal training result as a DCBN cable fault diagnosis model, takes the prediction result of each characteristic parameter of the high-voltage cable obtained by the decomposition module as the input of the DCBN cable fault diagnosis model, and realizes the recent partial discharge fault prediction of the cable.

Compared with the prior art, the invention has at least the following beneficial effects:

a high-voltage cable partial discharge fault prediction method utilizes high-voltage cable historical data and online data to realize the positioning and prediction diagnosis of a partial discharge fault, performs wavelet denoising and Variation Modal Decomposition (VMD) on an online high-frequency signal, selects the phase response at the highest amplitude in an IMF1, and combines cable double-end signals to realize the partial discharge positioning. Based on a multiple linear regression algorithm (MLR), the historical data is utilized to approach the correlation characteristics among the grounding current, the temperature, the load current and the PRPD map form statistical characteristic quantity, so that the real-time filling of the missing items in the online monitoring data is realized; based on a time sequence decomposition model (STL) and a time sequence prediction model (LSTM), realizing joint time sequence prediction of various parameters; and establishing a DCBN deep convolution confidence network, and mining potential association of various parameters and cable partial discharge types. Realizing the predictive positioning and diagnosis of the partial discharge faults of the high-voltage cable

Furthermore, in the long-term operation and maintenance process of the cable, a large amount of historical effective data is precipitated and collected into a cable partial discharge fault positioning data set, a cable state prediction historical data set and an original cable fault diagnosis historical data set can provide data support for cable fault positioning, state prediction and insulation fault diagnosis model training.

Furthermore, accurate partial discharge fault location can effectively reduce manpower and material resources consumed in the operation and maintenance process of the high-voltage cable. The cable high-frequency signal is subjected to modal component extraction, lagrange multiplier method and penalty factor alpha are introduced to enable partial discharge high-frequency signal variation decomposition to be restrained, a primary limited bandwidth natural modal function of the two high-frequency signals is obtained, and signal reference is provided for double-end phase response positioning.

Furthermore, the current research on the phase attenuation characteristic of the cable partial discharge is mature, and the wavelet analysis denoising and variation modal analysis of the current main stream can effectively acquire the phase response from the original high-frequency signal. And carrying out fast Fourier change on the original high frequency to obtain a sensor phase response map, combining the peak position of an intrinsic mode function (intrinsic mode function) IMF1 obtained after the original signal filtering and the variation mode decomposition (Variational Mode Decomposition) VMD to obtain the phase response of the sensor, and combining the phase attenuation characteristic equation to obtain the partial discharge fault position location.

Further, the physical quantity of the map statistics characteristic in the cable partial discharge PRPD map is extracted, and the key information in the partial discharge map is represented by a plurality of digital characteristic quantities, so that the subsequent data missing and time sequence are facilitatedPredicting; the specific physical quantities of the map statistical characteristics are as follows: skew S _k Steepness K _u The discharge factor Q, the correlation coefficient CC and the phase asymmetry psi. Inclination S _k Representing the degree of shape symmetry inclination of PRPD pattern, if S _k If equal to 0, the pattern is symmetrical, if S _k If the value is greater than 0, the shape of the map is inclined to be smaller than the arithmetic mean, if S _k A plot shape that is less than 0 indicates a slope of the plot shape that is greater than the arithmetic mean. Steepness K _u Representing the degree of protrusion of the PRPD pattern relative to normal distribution, if K _u If equal to 0, the shape of the map and normal distribution are generally smooth, if K _u If the distribution is larger than 0, the distribution is more sharp, if K _u Smaller than 0 then the flatter distribution is. The discharge factor Q represents the discharge variability over the positive and negative half periods in the PRPD pattern. The correlation coefficient CC represents the degree of similarity of the PRPD pattern contour over the positive and negative half periods, from 0 to 1, meaning that the contour difference is large to small. The phase asymmetry ψ represents the difference in the initial discharge phases of the positive and negative half periods of the PRPD pattern.

Further, the partial discharge corresponding to the high-voltage cable is a process of generating breakdown discharge in a partial area between voltage conductors applied in the operation process of the electric equipment, the type and the degree of the activity of the partial discharge necessarily cause different degrees of changes of the electric heating characteristics of the cable, the load current, the cable temperature and the grounding current are reflected by the electric heating characteristics of the cable, and the statistical characteristic physical quantity of the PRPD map is used as the statistical representation of the electric partial discharge degree and the phenomenon of the cable, so that strong correlation is necessarily generated between the partial discharge and the phenomenon. And utilizing historical data, mining potential relations among the multiple linear regression (multiple linear regression) MLR algorithms, realizing self-adaption to make up for data loss caused by different sensor sampling frequencies or sampling abnormality and other reasons in the cable monitoring system, and laying a foundation for the follow-up prediction.

Further, the seasonal trend decomposition method STL is a time series decomposition method which is widely used and has strong robustness, and can decompose a time series into a trend term, a period term and a remainder. And a Long Short-Term Memory (LSTM) utilizes the structures of an input gate, a forgetting gate and an output gate, so that the association among multiple parameters and the time sequence characteristics of data can be effectively mined, and the time sequence prediction of the multiple parameters is further realized. In order to achieve a higher-precision prediction effect, a trend term obtained after STL decomposition is used as an input of LSTM, and a period term obtained after STL decomposition is added with the prediction output of LSTM.

Further, the deep convolutional confidence network (Deep Convolution Belief Networks) DCBN is a deep learning classification method based on a large training dataset. The input data features are extracted by utilizing a convolution layer, feature extraction and conversion are realized by utilizing a plurality of layers of limited Boltzmann machines (Restricted Boltzmann Machine) RBM, and finally classification diagnosis is realized by using a fully connected perceptron (Multi layer Perceptron) MLP and a ReLU function (Rectified Linear Units) as activation functions. And finally, fine tuning the whole network by using Back Propagation (BP). The deep convolution confidence network has higher accuracy than a general classifier, so that the DCBN network is selected as a network template for the partial discharge fault diagnosis of the high-voltage cable.

Furthermore, by using the deep convolution confidence network (Deep Convolution Belief Networks) DCBN as a terminal model for cable fault diagnosis, the utilization rate of data and the timeliness of fault diagnosis can be effectively improved by the characteristics of deep mining and convenience of the deep convolution confidence network on input features, so that the rapid and accurate diagnosis of cable partial discharge fault prediction can be realized.

It will be appreciated that the advantages of the second aspect may be found in the relevant description of the first aspect, and will not be described in detail herein.

In conclusion, the invention realizes the positioning and predictive diagnosis analysis of the partial discharge faults of the high-voltage cable of the power grid, and can effectively sense the state of the insulation fault of the high-voltage cable, adaptively monitor and regulate energy and other capabilities.

The technical scheme of the invention is further described in detail through the drawings and the embodiments.

Drawings

FIG. 1 is a schematic diagram of the whole process of the state holographic sensing data driven high-voltage cable partial discharge fault prediction;

FIG. 2 is a flow chart of the present invention for online localization of high voltage cable partial discharge faults;

FIG. 3 is a flow chart of the present invention for implementing online ischemia based on multiple linear regression MLR;

FIG. 4 is a flow chart of the present invention for implementing multi-parameter joint timing prediction based on STL and LSTM;

FIG. 5 is a flow chart of the high voltage cable insulation fault diagnosis DCBN network model set up and training of the present invention;

FIG. 6 shows a two-terminal filtered high frequency signal obtained by experimental monitoring on the A-phase circuit of the cable of the present invention;

fig. 7 shows another two-terminal filtered high frequency signal obtained by experimental monitoring on the a-phase circuit of the cable of the present invention.

Detailed Description

The following description of the embodiments of the present invention will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are some, but not all embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

In the description of the present invention, it will be understood that the terms "comprises" and "comprising," when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.

It is also to be understood that the terminology used in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used in this specification and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.

It should be further understood that the term "and/or" as used in the present specification and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes such combinations, e.g., a and/or B, may represent: a exists alone, A and B exist together, and B exists alone. In addition, the character "/" herein generally indicates that the front and rear associated objects are an "or" relationship.

It should be understood that although the terms first, second, third, etc. may be used to describe the preset ranges, etc. in the embodiments of the present invention, these preset ranges should not be limited to these terms. These terms are only used to distinguish one preset range from another. For example, a first preset range may also be referred to as a second preset range, and similarly, a second preset range may also be referred to as a first preset range without departing from the scope of embodiments of the present invention.

Depending on the context, the word "if" as used herein may be interpreted as "at … …" or "at … …" or "in response to a determination" or "in response to detection". Similarly, the phrase "if determined" or "if detected (stated condition or event)" may be interpreted as "when determined" or "in response to determination" or "when detected (stated condition or event)" or "in response to detection (stated condition or event), depending on the context.

Various structural schematic diagrams according to the disclosed embodiments of the present invention are shown in the accompanying drawings. The figures are not drawn to scale, wherein certain details are exaggerated for clarity of presentation and may have been omitted. The shapes of the various regions, layers and their relative sizes, positional relationships shown in the drawings are merely exemplary, may in practice deviate due to manufacturing tolerances or technical limitations, and one skilled in the art may additionally design regions/layers having different shapes, sizes, relative positions as actually required.

The invention provides a high-voltage cable partial discharge fault prediction method, which is used for collecting multiple partial discharge data, temperature data and power data of on-line monitoring of a cable and realizing the partial discharge positioning of the on-line monitoring through filtering impurity removal and modal extraction; establishing a historical fault data set of the high-voltage cable, and training to obtain a partial discharge diagnosis DCBN model; extracting morphological statistical feature quantity in an online acquisition PRPD map; filling the missing items of the rest monitoring parameters of the time sequence missing; respectively extracting each trend item and each period item by using the rest monitoring parameters and the extracted partial discharge morphological statistical characteristic quantity, and then combining and constructing a trend prediction model; and taking the trend prediction results of the rest monitoring parameters and the partial discharge form statistics characteristic quantity as the input of a partial discharge diagnosis model to obtain a high-voltage cable partial discharge fault diagnosis result. The prediction, positioning and diagnosis of the high-voltage cable partial discharge faults are realized. Based on historical operation data and on-line monitoring data, the invention realizes real-time positioning and auxiliary decision-making of the insulation operation risk of the high-voltage cable, and is beneficial to improving the reliability early warning, global perception, strong robust regulation and control and other multi-aspect capabilities of the digital twin of the power system.

Referring to fig. 1, the prediction method of the present invention is divided into three parts, namely three data acquisition, online positioning and prediction diagnosis, and based on the phase attenuation characteristic of the cable partial discharge high-frequency signal in the propagation process, the online positioning of the partial discharge fault is realized by using the online acquisition high-frequency signal. Utilizing a multiple linear regression method, according to the complete time node data set, arranging potential relations among the fitting parameters, realizing missing items in the sequence when the historical data are complemented, and laying a rammed data foundation for subsequent prediction; and then based on the STL and LSTM models, the multi-parameter joint time sequence prediction is realized by comprehensively mining the development trend and the period quantity in the historical data. And finally, based on the DCBN network, building a partial discharge fault diagnosis model by taking PRPD pattern form statistics feature quantity, ground current, load current and light measurement temperature as input features, and realizing final high-voltage cable insulation fault prediction diagnosis.

The data acquisition is divided into two types, one type is a partial discharge monitoring signal, and the other type is the rest monitoring parameters;

the partial discharge monitoring signal comprises a high-frequency signal and a PRPD map; the other monitoring parameters are ground current, load current and light temperature.

The specific sampling period is flexibly adjusted by a sensor in the field cable monitoring system.

The on-line positioning utilizes high-frequency signals in the partial discharge monitoring signals, and a high-frequency signal group acquired by two adjacent high-frequency sensors is needed;

using classical wavelet analysis method, selecting wavelet base and threshold value, noise reduction and purification are carried out to original high-frequency signal;

then, performing fast Fourier transform on the acquired original signals to obtain respective phase response maps of the two sensor signals;

extracting a finite bandwidth intrinsic mode function IMF1 of the high-frequency signal after noise reduction by using a variational mode decomposition method, and combining a phase response map and an IMF1 peak position to obtain phase response of the two high-frequency signals

And->

Finally, the phase attenuation characteristic of the partial discharge of the cable is utilized to be combined

And->

And obtaining the specific position of the partial discharge fault point between the two high-frequency sensors.

The data required for predictive diagnosis are PRPD pattern, ground current, load current and photometric temperature. The operation process can be divided into four parts: PRPD map morphology statistics feature extraction, data deficiency and supplement, time sequence prediction and fault diagnosis. The PRPD pattern morphology statistical feature quantity extraction is to extract physical quantities which can effectively represent statistical physical features in a pattern, and comprises the following steps: skew S _k Steepness K _u The discharge factor Q, the correlation coefficient CC and the phase asymmetry psi.

The data missing supplement is to supplement characteristic parameter data of missing time nodes by using multiple linear regression, and aims to prevent data asymmetry or missing caused by different sensor sampling frequencies or equipment anomalies. And (3) utilizing the complete data set in the historical time sequence data, taking the missing item as a dependent variable, taking the non-missing item as an independent variable, and utilizing the historical data to approximate the relation between the independent variable and the dependent variable. In order to achieve excellent timeliness in field application and achieve rapid predictive diagnosis, various characteristic parameters can be arranged and combined to obtain a perfect data missing-compensation model, and missing items are automatically compensated in the process of collecting data by a sensor and uploading the data to a database.

The time sequence prediction is to utilize STL and LSTM to realize trend prediction of various future characteristic parameters of the high-voltage cable based on complete multi-parameter historical time sequence data. And (3) performing component decomposition on the historical time sequence data of each parameter by using STL, and retaining trend items and period items therein. And (3) utilizing the LSTM, taking the window length of which the period length of the period item with the optimal LSTM prediction result in the STL decomposition result is the LSTM, realizing multi-parameter period item joint time sequence prediction, and finally adding respective period items to each parameter to obtain the final prediction result.

The fault diagnosis is to obtain a final high-voltage cable insulation fault prediction diagnosis result by using a DCBN neural network model and taking a plurality of parameters as input characteristics of the DCBN based on a historical fault data set, wherein the final high-voltage cable insulation fault prediction diagnosis result comprises the establishment and training of the model and the use of the model. For the establishment and training of the DCBN model, the number of neurons of an input layer and an output layer of a network is required to be determined according to the quantity of parameters and the possible type number of cable insulation faults, the number of the RBM layers, the number of neurons of each RBM layer, the parameters of a convolution layer, parameters of a pooling layer, parameters of convolution kernels, parameters of packing and the like are determined through parameter tuning and optimizing, and finally the DCBN model with the optimal effect is stored. The model is used, the prediction results of a plurality of parameters are used as the stored DCBN model input, and the final prediction diagnosis result is obtained according to the probability of various fault types output by the model.

Referring to fig. 2, the method for predicting the partial discharge fault of the high-voltage cable of the present invention includes the following steps:

s1, acquiring required cable state data by utilizing an online monitoring device, wherein the cable state data comprise a PRPD map, a high-frequency signal, a grounding current, a temperature and a load current; the high-frequency signals collected by two high-frequency sensors adjacent to each other in the position on the cable are taken as a group to be used as a cable partial discharge fault positioning data set; taking the PRPD map, the grounding current, the temperature and the load current as a cable state prediction historical data set; collecting the normal state of the cable and generating partial laid PRPD map, grounding current, temperature and load current data to form an original cable fault diagnosis historical data set;

S2, obtaining two high-frequency signals in the cable partial discharge fault location data set by utilizing a wavelet filtering processing step S1, carrying out wavelet decomposition on the original signals to obtain scale coefficients, screening out high-frequency noise through threshold processing, and reversely reconstructing to obtain filtered signals;

s201, selecting a wavelet basis, and considering that a partial discharge signal is exponentially attenuated and oscillated, selecting a similar wavelet basis to facilitate denoising;

s202, determining a decomposition scale, wherein the decomposition scale is used for achieving the purpose of signal-to-noise separation and preventing distortion;

s203, defining a threshold value and a threshold function;

s204, reversely reconstructing to obtain the purified high-frequency signal.

S3, respectively extracting modal components of the two high-frequency signals subjected to data processing obtained in the step S2, enabling the partial discharge high-frequency signals to be subjected to variational decomposition to have constraint by introducing Lagrange multiplier method and penalty factor alpha, and performing iterative optimization by using Lagrange multiplier factor and update factor sigma through an alternate direction algorithm until convergence standard fault tolerance epsilon is met, so as to obtain a series of finite bandwidth Intrinsic Mode Functions (IMFs) after respective decomposition, and finally obtaining primary finite bandwidth intrinsic mode functions (IMF 1) of the two high-frequency signals;

s4, based on a double-end phase response method, processing two high-frequency signals in the cable partial discharge fault location data set in the step S1 by utilizing fast Fourier distribution change to obtain respective phase response maps, combining phase responses of peaks in the two IMFs 1 obtained in the step S3, combining the distance between the two sensors, and utilizing a cable partial discharge phase attenuation characteristic equation to realize fault location of cable partial discharge;

The fault location realized by the double-end phase response method is specifically as follows:

s401, acquiring two high-frequency sensor signals adjacent to each other on a cable;

s402, performing fast Fourier transform on the acquired signals to respectively obtain two phase response maps;

s403, obtaining IMF1 of the step S2 and the step S3;

s404, respectively acquiring phase responses of peak values in the two IMFs 1

And->

S405, simultaneous

And->

The location of the partial discharge fault is determined using the following formula.

Wherein, the liquid crystal display device comprises a liquid crystal display device,

the peak phase of the original partial discharge high-frequency signal, d is the partial discharge position, l is the distance between two sensors, beta ₁ And beta ₂ The phase constants of the partial discharge signals of the cable positions where the two sensors are located are calculated by the following formula:

wherein R is ₀ 、G ₀ 、L ₀ And C ₀ The resistance, conductance, inductance and capacitance values of the cables per unit length, respectively, ω being the angular frequency of the peak position of the partial discharge IMF 1.

S5, extracting the morphological statistical feature quantity of the PRPD map in the cable state prediction historical data set obtained in the step S1, and replacing the PRPD map with the morphological statistical feature quantity of the PRPD map to form a new cable state prediction historical data set;

PRPD profile feature extraction includes the following: skew S _k Steepness K _u The discharge factor Q, the correlation coefficient CC and the phase asymmetry psi;

skew S _k The calculation method of (2) is shown in the formula (3):

wherein N is the number of windows in the half power frequency period of the map,

is the phase of the i-th phase window, +.>

Is the phase width, mu, p _i And sigma are respectively +.>

The calculation formulas are shown in formulas (4) to (6) for the mean value, probability and variance of the occurrence of the i-th phase window partial discharge of the spectrum in the variable time, wherein y is the ordinate of the two-dimensional spectrum,

steepness K _u The calculation formula of (2) is as follows:

the formula for calculating the discharge factor Q is:

and->

Sum of discharge amounts of positive half period and negative half period of phase, respectively, +.>

And->

The sum of the discharge times of the positive half cycle and the negative half cycle respectively;

the calculation formula of the correlation coefficient CC is:

and->

Average discharge amounts of a positive half period and a negative half period of an ith phase window of the map respectively; the calculation formula of the phase asymmetry ψ is:

and->

The pattern initiates discharge phases in the positive and negative half-cycles, respectively.

S6, filling the data of each cable characteristic parameter in the cable state prediction historical data set obtained in the step S5 by utilizing multiple linear regression, and filling the missing characteristic parameters, so that the cable state history data set is formed, and reliable data support is provided for subsequent time sequence prediction;

Referring to fig. 3, the steps of data missing interpolation of multiple linear regression are specifically:

s602, using the missing item as a dependent variable and the non-missing item as an independent variable, approximating the following formula by using the historical data:

αβ+ε＝ζ (11)

where α is the argument, ζ is the argument, β is the weight matrix, and ε is the bias matrix.

S603, filling the missing items by using the fitting result.

S7, decomposing each item of historical data in the cable state prediction historical data set obtained in the step S6 into a Trend Component (Trend Component), a period Component (Seasonal Component) and a remainder (Remainder Component) by using a time sequence Trend decomposition algorithm STL (Seaseal-Trend decomposition procedure based on Loess), and realizing joint time sequence prediction of each item of historical data Trend Component through a Long Short-Term Memory (LSTM) to obtain a prediction result of each characteristic parameter of the cable;

referring to fig. 4, the combined time sequence prediction of the historical data of each characteristic parameter of the cable by using the time sequence decomposition STL and the long short term memory network LSTM is specifically as follows:

s702, defining the required prediction time span;

s703, taking the time span of the period amount obtained by STL decomposition as the length of an input window of LSTM (for example, the STL is decomposed into a plurality of different time spans, and selecting the optimal prediction effect to obtain one;

and S705, adding trend prediction results of different parameters with respective period components to realize multi-parameter joint timing prediction.

S8, extracting morphological statistical feature quantity of PRPD (partial pressure differential) map in cable fault diagnosis historical data set obtained in step S1, wherein the morphological statistical feature quantity comprises deflection S _k Steepness K _u The discharge factor Q, the correlation coefficient CC and the phase asymmetry psi are utilized to replace the PRPD by using the morphological statistical feature quantity of the PRPD to form a new cable fault diagnosis historical data set;

s9, establishing a fault diagnosis model of a deep convolution confidence network (DCBN) by using the cable fault diagnosis historical data set obtained in the step S8, and storing a model with an optimal training result as the DCBN cable fault diagnosis model. And (3) taking the prediction results of all characteristic parameters of the cable obtained in the step (S7) as the input of a DCBN cable fault diagnosis model to realize the prediction of the recent partial discharge fault of the cable.

Referring to fig. 5, the overall flow of the high voltage cable insulation fault diagnosis DCBN network model setup and training is as follows:

s901, acquiring a historical fault data set, and then going to step S902;

s902, determining the number of neurons of an input layer and an output layer, the number of layers of RBM of a network structure, the number of neurons of each layer of RBM, parameters of a convolution layer and a pooling layer, convolution kernel, padding and the like, and then going to step S903;

s903, performing one-hot coding on the partial discharge fault type of the data set as output of the DCBN network, and then going to step S904;

s904, normalizing the characteristic parameters to [ -1,1] and dividing a training set and a testing set, and then going to a step S905;

s905, training by using the training set and optimizing by using the test set, and then going to step S906;

s906, storing a model with the optimal test set effect as a DCBN cable fault diagnosis model.

Referring to fig. 6, fault diagnosis using the deep convolutional confidence network DCBN is specifically:

s907, normalizing the cable fault diagnosis historical data set obtained in the step S8 to form [ -1,1] as an input of a DCBN network, and performing one-hot coding on the partial discharge type as an output;

s908, setting interface parameters of DCBN, namely the number of neurons of input layers and output layers, wherein the number of the input layers is equal to the number of characteristic parameters, and the number of the output layers is equal to the number of partial discharge types;

S909, setting a DCBN internal network structure, namely the number of RBM layers, the number of neurons of each RBM layer, parameters of a convolution layer and a pooling layer, convolution kernel, padding and the like;

s910, extracting an output effective characteristic value by utilizing CNN of the front end of the network;

s911, performing unsupervised optimization on the multi-layer RBM of the DCBN;

s912, optimizing the network overall situation by using a BP algorithm by utilizing the full connection of the network terminal;

s913, repeatedly training the network by using the data set normalized in the step S907, and storing a model with the best diagnosis effect;

s914, normalizing the multi-parameter joint time sequence prediction result obtained in the step S7 by using the normalization scale of the historical data set in the step S907;

s915, inputting the result of the step S914 to the DCBN to obtain a final diagnosis result, and realizing cable partial discharge fault prediction.

In still another embodiment of the present invention, a high-voltage cable partial discharge fault prediction system is provided, where the system can be used to implement the high-voltage cable partial discharge fault prediction method described above, and specifically, the high-voltage cable partial discharge fault prediction system includes a data module, a reconstruction module, an extraction module, a positioning module, a first replacement module, a filling module, a decomposition module, a second replacement module, and a prediction module.

The data module is used for collecting state data of the high-voltage cable and constructing a cable partial discharge fault positioning data set, a cable state prediction historical data set and an original cable fault diagnosis historical data set;

In yet another embodiment of the present invention, a terminal device is provided, the terminal device including a processor and a memory, the memory for storing a computer program, the computer program including program instructions, the processor for executing the program instructions stored by the computer storage medium. The processor may be a central processing unit (Central Processing Unit, CPU), but may also be other general purpose processor, digital signal processor (Digital Signal Processor, DSP), application specific integrated circuit (Application Specific Integrated Circuit, ASIC), off-the-shelf programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic device, discrete gate or transistor logic device, discrete hardware components, etc., which are the computational core and control core of the terminal adapted to implement one or more instructions, in particular to load and execute one or more instructions to implement the corresponding method flow or corresponding functions; the processor of the embodiment of the invention can be used for the operation of the high-voltage cable partial discharge fault prediction method, and comprises the following steps:

collecting state data of a high-voltage cable, constructing a cable partial discharge fault positioning data set, a cable state prediction historical data set and an original cable fault diagnosis historical data set; processing the high-frequency signals in the cable partial discharge fault positioning data set by utilizing wavelet filtering, and reversely reconstructing to obtain filtered high-frequency signals; extracting modal components of the high-frequency signal to obtain a primary limited bandwidth inherent modal function of the high-frequency signal; based on a double-end phase response method, processing high-frequency signals in the cable partial discharge fault location data set by utilizing fast Fourier distribution change to obtain a corresponding phase response map, and simultaneously setting up phase response of peak values in the high-frequency signals in the first finite bandwidth inherent mode function, and realizing fault location of the high-voltage cable partial discharge by utilizing a cable partial discharge phase attenuation characteristic equation; extracting morphological statistical feature quantity of the PRPD map in the cable state prediction historical data set, and replacing the PRPD map by using the morphological statistical feature quantity of the PRPD map to form a new cable state prediction historical data set; filling data of each cable characteristic parameter in the new cable state prediction historical data set by utilizing multiple linear regression to form a cable state historical data set; decomposing each item of historical data in the cable state historical data set into a trend component, a period component and a remainder by using a time sequence trend decomposition algorithm, and then realizing the combined time sequence prediction of each item of historical data trend component by using a long-short-period memory network to obtain a prediction result of each item of characteristic parameters of the high-voltage cable; extracting morphological statistical feature quantity of the PRPD map in the cable fault diagnosis historical data set, and replacing the PRPD map by using the morphological statistical feature quantity of the PRPD map to form a new cable fault diagnosis historical data set; and establishing a fault diagnosis model of the deep convolution confidence network by using the new cable fault diagnosis historical data set, storing a model with an optimal training result as a DCBN cable fault diagnosis model, and taking the prediction result of each characteristic parameter of the high-voltage cable as the input of the DCBN cable fault diagnosis model to realize the prediction of the recent partial discharge fault of the cable.

In a further embodiment of the present invention, the present invention also provides a storage medium, in particular, a computer readable storage medium (Memory), which is a Memory device in a terminal device, for storing programs and data. It will be appreciated that the computer readable storage medium herein may include both a built-in storage medium in the terminal device and an extended storage medium supported by the terminal device. The computer-readable storage medium provides a storage space storing an operating system of the terminal. Also stored in the memory space are one or more instructions, which may be one or more computer programs (including program code), adapted to be loaded and executed by the processor. The computer readable storage medium may be a high-speed RAM Memory or a Non-Volatile Memory (Non-Volatile Memory), such as at least one magnetic disk Memory.

One or more instructions stored in a computer-readable storage medium may be loaded and executed by a processor to implement the respective steps of the high-voltage cable partial discharge fault prediction method in the above embodiments; one or more instructions in a computer-readable storage medium are loaded by a processor and perform the steps of:

For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention. The components of the embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations. Thus, the following detailed description of the embodiments of the invention, as presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

The present invention will be described with reference to experimental data for a 220kv three-phase cable. Fig. 6 and 7 show the high frequency signals obtained by filtering the two ends of the cable a phase circuit obtained by monitoring a certain experiment. According to the high-voltage cable partial discharge fault prediction method and system provided by the invention, the original signals at two ends are subjected to fast Fourier transform and modal decomposition, so that two IMF1 peak phase responses are respectively 1.37rad and 1.25rad. The simultaneous equations 1 and 2 determine the partial discharge position with a final positioning error of 0.023%.

The load current, photodetection temperature and a-phase ground current history of the cable are shown in the table.

Table 1 cable history data

The load current, the photodetection temperature and the A-phase grounding current are used as prediction objects, the MLR is utilized to complement the PRPD pattern form statistical characteristic quantity, and then a prediction model is established based on the STL-LSTM model. And (3) debugging and selecting a prediction model with the best fitting degree for multiple times, wherein the fitting accuracy rates of the prediction model on load current, photodetection temperature and A-phase grounding current are 93.235%, 95.856% and 96.519%, respectively. Thus, STL-LSTMLSTM may be used as a base model for predicting cable status. The predicted results of the cable for the next 2 days are shown in the table.

Table 2 cable forecast data

Prediction time	Load current/A	Photodetection temperature/°c	Ground current/A
				2022-08-19	272	35.6	6.95
2022-08-20	265	35.8	6.86

And establishing a DCBN-based high-voltage cable partial discharge fault diagnosis model according to the historical fault data set. 8:2, dividing the historical fault data set into a training set and a testing set, and after model parameter optimization is carried out through the DCBN, respectively obtaining the model parameters of the DCBN: the number of hidden layers is 8, the rbm learning rate is 0.1, the activation function is ReLU, and the padding is 2. The result shows that the diagnosis accuracy of the training set is 96.568%, the diagnosis accuracy of the testing set is 94.056%, and the established DCBN fault diagnosis model can realize effective detection of cable fault diagnosis by utilizing a plurality of monitoring signals of the cable. Therefore, the prediction results of table 2 are diagnosed, and the diagnosis results are shown in table 3, and the probability of the occurrence of the creeping discharge fault is highest in two days in the future, and the fault is the air gap discharge fault.

TABLE 3 predictive diagnosis of Cable faults

The invention has the following characteristics:

(1) The high-voltage cable online positioning and fault prediction method is combined and utilized, the sequence of positioning and then diagnosing is followed, and the online supervision, the rapid positioning, the trend prediction and the fault pre-diagnosis of the high-voltage cable can be realized. Not only can provide indication reference for reliable operation and maintenance of the high-voltage cable, but also can effectively save required time and resources.

(2) By extracting the statistical characteristic physical quantity of the PRPD map of the partial discharge of the high-voltage cable, the data is reserved by using smaller storage resources, and convenience is brought to data defect compensation and prediction. The method reduces the calculation amount of the program, saves time and improves precision.

(3) The PRPD map statistical characteristic physical quantity, load current, cable temperature and grounding current are taken as characteristic data, the potential relation among the PRPD map statistical characteristic physical quantity, the load current, the cable temperature and the grounding current is mined by using an MLR algorithm, and a data missing compensation model is established, so that the grounding self-adaptability of the whole scheme is improved, and the PRPD map statistical characteristic physical quantity, the load current, the cable temperature and the grounding current are not limited by conditions such as asynchronous sampling frequency, instability and the like in a high-voltage cable monitoring system.

(4) By combining STL and LSTM algorithms, reliable time sequence prediction of PRPD map statistical characteristic physical quantity, load current, cable temperature and grounding current is realized, and characteristic sources are provided for high-voltage cable fault prediction diagnosis, so that the high-voltage cable fault prediction diagnosis is possible. The method has important significance for guiding the operation and maintenance of the high-voltage cable aiming at the insulation fault.

In summary, according to the method and the system for predicting the partial discharge fault of the high-voltage cable, the partial discharge position is positioned by utilizing the phase attenuation characteristic of the cable partial discharge high-frequency signal in the propagation process. In order to lay a rammed data foundation for time sequence prediction, a multiple linear regression method is utilized to complement missing items in the sequence when required data is obtained. The method has the advantages that the on-line available signals of the high-voltage cable are utilized to jointly excavate trend movement in data time and the association of data structure components and fault types based on a time sequence decomposition (STL) method and a long and short term memory network (LSTM), the prediction diagnosis is realized on the insulation faults of the high-voltage cable, the historical precipitation data of the cable can be effectively utilized, the potential value is excavated, the partial discharge fault positioning and the prediction diagnosis of the high-voltage cable are realized, and a certain guiding effect is realized on the operation and maintenance of the high-voltage cable.

It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-described division of the functional units and modules is illustrated, and in practical application, the above-described functional distribution may be performed by different functional units and modules according to needs, i.e. the internal structure of the apparatus is divided into different functional units or modules to perform all or part of the above-described functions. The functional units and modules in the embodiment may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit, where the integrated units may be implemented in a form of hardware or a form of a software functional unit. In addition, specific names of the functional units and modules are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present application. The specific working process of the units and modules in the above system may refer to the corresponding process in the foregoing method embodiment, which is not described herein again.

In the foregoing embodiments, the descriptions of the embodiments are emphasized, and in part, not described or illustrated in any particular embodiment, reference is made to the related descriptions of other embodiments.

Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.

In the embodiments provided in the present invention, it should be understood that the disclosed apparatus/terminal and method may be implemented in other manners. For example, the apparatus/terminal embodiments described above are merely illustrative, e.g., the division of the modules or units is merely a logical function division, and there may be additional divisions when actually implemented, e.g., multiple units or components may be combined or integrated into another system, or some features may be omitted or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection via interfaces, devices or units, which may be in electrical, mechanical or other forms.

The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

In addition, each functional unit in the embodiments of the present invention may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.

The integrated modules/units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the present invention may implement all or part of the flow of the method of the above embodiment, or may be implemented by a computer program to instruct related hardware, where the computer program may be stored in a computer readable storage medium, and when the computer program is executed by a processor, the computer program may implement the steps of each of the method embodiments described above. Wherein the computer program comprises computer program code which may be in source code form, object code form, executable file or some intermediate form etc. The computer readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a U disk, a removable hard disk, a magnetic disk, an optical disk, a computer Memory, a Read-Only Memory (ROM), a random access Memory (RandomAccess Memory, RAM), an electrical carrier wave signal, a telecommunications signal, a software distribution medium, etc., it should be noted that the computer readable medium may contain content that is appropriately increased or decreased according to the requirements of jurisdictions and patent practices, such as in certain jurisdictions, according to the jurisdictions and patent practices, the computer readable medium does not contain electrical carrier wave signals and telecommunications signals.

The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

The above is only for illustrating the technical idea of the present invention, and the protection scope of the present invention is not limited by this, and any modification made on the basis of the technical scheme according to the technical idea of the present invention falls within the protection scope of the claims of the present invention.

Claims

1. The high-voltage cable partial discharge fault prediction method is characterized by comprising the following steps of:

2. The method for predicting partial discharge faults of a high-voltage cable according to claim 1, wherein in step S1, PRPD map, high-frequency signal, ground current, temperature and load current data of the high-voltage cable are collected; taking high-frequency signals acquired by two high-frequency sensors adjacent to each other on a high-voltage cable as a cable partial discharge fault positioning data set; taking the PRPD map, the grounding current, the temperature and the load current as a cable state prediction historical data set; and collecting the data of the PRPD map, the grounding current, the temperature and the load current of the high-voltage cable in normal state and the generation of partial drop, and forming an original cable fault diagnosis historical data set.

3. The method for predicting partial discharge faults of high-voltage cables according to claim 1, wherein in the step S3, lagrange multiplier method and penalty factor alpha are introduced to make partial discharge high-frequency signals variational decomposition have constraint, lagrange multiplication factor and update factor sigma are used for iterative optimization through an alternate direction algorithm until convergence standard fault tolerance epsilon is met, a series of finite bandwidth intrinsic mode functions after respective decomposition are obtained, and finally primary finite bandwidth intrinsic mode functions of two high-frequency signals are obtained.

4. The method for predicting partial discharge failure of a high-voltage cable according to claim 1, wherein step S4 specifically comprises:

And->

Simultaneous->

And->

And determining the position of the partial discharge fault.

5. The method according to claim 1, wherein in step S5, the PRPD pattern feature extraction includes a skew S _k Steepness K _u A discharge factor Q, a correlation coefficient CC, and a phase asymmetry ψ.

6. The method for predicting partial discharge failure of a high voltage cable according to claim 1, wherein in step S6, the data missing supplements of multiple linear regression are specifically:

s603, filling the missing items by using the fitting result.

7. The method for predicting partial discharge failure of a high voltage cable according to claim 1, wherein in step S7, the performing of the joint timing prediction specifically includes:

s702, definitely needing predicted time span;

8. The method for predicting partial discharge faults of high voltage cables according to claim 1, wherein in step S9, fault diagnosis by using a deep convolutional confidence network DCBN is specifically:

9. The high-voltage cable partial discharge fault prediction method according to claim 8, wherein the optimal DCBN cable fault diagnosis model is specifically:

10. A high voltage cable partial discharge fault prediction system, comprising: