CN106778692B - Cable partial discharge signal identification method and device based on S transformation - Google Patents
Cable partial discharge signal identification method and device based on S transformation Download PDFInfo
- Publication number
- CN106778692B CN106778692B CN201710035996.3A CN201710035996A CN106778692B CN 106778692 B CN106778692 B CN 106778692B CN 201710035996 A CN201710035996 A CN 201710035996A CN 106778692 B CN106778692 B CN 106778692B
- Authority
- CN
- China
- Prior art keywords
- partial discharge
- discharge signal
- singular value
- matrix
- module
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
- G06F2218/16—Classification; Matching by matching signal segments
- G06F2218/18—Classification; Matching by matching signal segments by plotting the signal segments against each other, e.g. analysing scattergrams
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Signal Processing (AREA)
- Testing Relating To Insulation (AREA)
Abstract
The embodiment of the invention discloses a cable partial discharge signal identification method and device based on S transformation, which are used for solving the technical problems that the existing method for positioning a power cable partial discharge signal source usually adopts a waveform time difference method, the identification precision is low and the required identification time is too long. The method provided by the embodiment of the invention comprises the following steps: s transformation is carried out on the obtained partial discharge signals of known sources to obtain a complex time-frequency matrix; performing module solving on the complex time-frequency matrix to obtain a module matrix, and performing singular value decomposition on the module matrix to obtain a singular value sequence of the module matrix; equally dividing the singular value sequence into at least two intervals, calculating the ratio of the Shannon entropy of the singular value in each interval to the Shannon entropy of the singular value sequence, establishing a partial discharge signal characteristic sample library by taking the ratio as a partial discharge signal characteristic vector, and constructing a support vector machine model; and inputting the characteristic vector of the partial discharge signal to be identified into the support vector machine model to obtain the source of the partial discharge signal to be identified.
Description
Technical Field
The invention relates to the technical field of cable partial discharge online monitoring, in particular to a cable partial discharge signal identification method and device based on S transformation.
Background
With the high-speed continuous development of power systems, the length of a laying loop of a power cable is steadily increased, and the power cable is widely applied to cities. However, the power load and voltage grade are increasing, and the insulation problem caused by the local defect of the cable poses great threats to the power supply quality, the social economy and the like. In order to monitor the insulation state of the cable and find local defects of the cable in time, thereby preventing the occurrence of cable operation accidents and ensuring the reliability of power grid operation, the local defects of the cable need to be detected.
In the cable partial discharge online monitoring, the detected partial discharge signal may come from the cable body and the cable terminal, and may also come from a switch cabinet connected with the cable body and the cable terminal. Because the partial discharge of different sources has different damages to equipment and different judgment standards, the method has important practical significance for identifying the source of the partial discharge signal.
In terms of partial discharge signal identification, signal feature extraction and classifier selection are the most critical parts. The feature extraction is the first step of partial discharge signal identification, and the quality of the feature extraction directly influences the identification effect. At present, the partial discharge signal feature extraction method mainly comprises two main categories of a statistical feature method and a time domain analysis method. The statistical characteristic method relates to the phase of partial discharge signals, while the distribution cable is generally a three-core cable and shares the same ground, and when partial discharge occurs in two phases or three phases, the detection of the phase characteristics of the partial discharge signals becomes almost impossible. The time domain analysis method is used for carrying out pattern recognition on waveform characteristics or corresponding transformation results obtained by collecting time domain pulses generated by one-time discharge at a high speed, and mainly comprises a Fourier analysis method, a wavelet analysis method, a waveform parameter direct extraction method and the like. The pattern recognition classifier mainly comprises a neural network classifier, a minimum distance classifier and a fuzzy recognition classifier. The neural network is easy to be converged in the defect of the local optimal solution, and the precision is not high.
At present, the whole power industry generally adopts a waveform time difference method for positioning a power cable partial discharge signal source, and the identification precision is greatly reduced and the required identification time is too long when a power cable line for transmitting the partial discharge signal is too long.
Disclosure of Invention
The embodiment of the invention provides a method and a device for identifying a cable partial discharge signal based on S transformation, which solve the technical problems that the identification precision is greatly reduced and the required identification time is too long when a power cable line for transmitting the partial discharge signal is too long because a waveform time difference method is usually adopted for positioning a power cable partial discharge signal source in the power industry at present.
The embodiment of the invention provides a cable partial discharge signal identification method based on S transformation, which comprises the following steps:
acquiring a partial discharge signal from a known source, and performing S transformation on the partial discharge signal to obtain a complex time-frequency matrix;
performing module solving on the complex time-frequency matrix to obtain a module matrix, and performing singular value decomposition on the module matrix to obtain a singular value sequence of the module matrix;
equally dividing the singular value sequence into at least two intervals according to the singular value sequence, calculating the ratio of the Shannon entropy of the singular value in each interval to the Shannon entropy of the singular value sequence, and establishing a partial discharge signal characteristic sample library by taking the ratio as a partial discharge signal characteristic vector;
taking the partial discharge signal characteristic vectors in the partial discharge signal characteristic sample library as input, and constructing a multi-classification support vector machine model;
taking the partial discharge signal as a sample, and training the support vector machine model to obtain a trained support vector machine model;
and inputting the characteristic vector of the partial discharge signal to be identified into a trained support vector machine model to obtain the source of the partial discharge signal to be identified.
Optionally, the partial discharge signal includes a cable body partial discharge signal, a cable termination partial discharge signal, a switch cabinet corona discharge signal, and a switch cabinet surface discharge signal.
Optionally, obtaining a partial discharge signal from a known source, and performing S-transform on the partial discharge signal to obtain a complex time-frequency matrix includes:
acquiring partial discharge signals from known sources, and performing S conversion on the partial discharge signals through a preset formula to obtain a complex time-frequency matrix, wherein the preset formula I specifically comprises the following steps:
where H (kt) is a discrete time sequence of the partial discharge signal, ST is a complex time-frequency matrix obtained by transforming the discrete time sequence of the partial discharge signal by S, T is a sampling period of the discrete time sequence, N is a length of the discrete time sequence, H is a fourier transform of the discrete time sequence, j is an imaginary unit, k, N, m is 0,1, …, N-1.
Optionally, performing modulo on the complex time-frequency matrix to obtain a mode matrix, and performing singular value decomposition on the mode matrix to obtain a singular value sequence of the mode matrix includes:
performing modulus calculation on the complex time-frequency matrix ST to obtain a modulus matrix STA, and performing singular value decomposition on the modulus matrix STA through a preset second formula to obtain a singular value sequence of the modulus matrix, wherein the preset second formula specifically comprises the following steps:
where U and V are both N × order N orthogonal matrices, and D ═ diag (σ)1,σ2,…,σN) Is a diagonal matrix whose diagonal elements (σ)1,σ2,…,σN) Is the singular value of the matrix STA, uiAnd viThe vectors of the singular values of the ith column of the matrices U and V, respectively.
Optionally, equally dividing the singular value sequence into at least two intervals according to the singular value sequence, calculating a ratio of Shannon entropy of the singular value in each interval to Shannon entropy of the singular value sequence, and establishing a partial discharge signal feature sample library by using the ratio as a partial discharge signal feature vector includes:
equally dividing the singular value sequence into at least two intervals according to the singular value sequence, calculating the ratio of the Shannon entropy of the singular value in each interval to the Shannon entropy of the singular value sequence, and establishing a partial discharge signal characteristic sample library by using the ratio as a partial discharge signal characteristic vector through a preset third formula, wherein the preset third formula specifically comprises the following steps:
λ＝[E1/E,E2/E,Eq/E…,EQ/E]；
Optionally, the source category of the partial discharge signal feature vector is marked by a two-bit binary number.
Optionally, the constructing a multi-classification support vector machine model with the partial discharge signal feature vectors in the partial discharge signal feature sample library as inputs includes:
and combining all the source categories of the partial discharge signal characteristic vectors in the partial discharge signal characteristic sample library by using a binary classification algorithm to form a plurality of sub-classifiers, and constructing a multi-classification support vector machine model by taking the partial discharge signal characteristic vectors in the partial discharge signal characteristic sample library as input.
Optionally, training the support vector machine model by using the partial discharge signal as a sample, and obtaining the trained support vector machine model includes:
randomly selecting partial discharge signals of different sources with the same quantity as training samples according to different sources of the partial discharge signals, and inputting the training samples into the support vector machine model for training to obtain the trained support vector machine model.
Optionally, inputting the feature vector of the partial discharge signal to be identified into the trained support vector machine model, and obtaining the source of the partial discharge signal to be identified includes:
and inputting the characteristic vector of the partial discharge signal to be identified into a trained support vector machine model to obtain an output value, and comparing the output value with the binary number label of the source category of the characteristic vector of the partial discharge signal to obtain the source of the partial discharge signal to be identified.
The embodiment of the invention provides a cable partial discharge signal identification device based on S transformation, which is characterized by comprising the following components:
the transformation module is used for acquiring partial discharge signals from known sources and performing S transformation on the partial discharge signals to obtain a complex time-frequency matrix;
the decomposition module is used for performing module solving on the complex time-frequency matrix to obtain a module matrix and performing singular value decomposition on the module matrix to obtain a singular value sequence of the module matrix;
the calculation module is used for equally dividing the singular value sequence into at least two intervals according to the singular value sequence, calculating the ratio of the Shannon entropy of the singular value in each interval to the Shannon entropy of the singular value sequence, and establishing a partial discharge signal characteristic sample library by taking the ratio as a partial discharge signal characteristic vector;
the building module is used for building a multi-classification support vector machine model by taking the partial discharge signal characteristic vectors in the partial discharge signal characteristic sample library as input;
the training module is used for training the support vector machine model by taking the partial discharge signal as a sample to obtain a trained support vector machine model;
and the input module is used for inputting the characteristic vector of the partial discharge signal to be identified into the trained support vector machine model to obtain the source of the partial discharge signal to be identified.
According to the technical scheme, the embodiment of the invention has the following advantages:
the embodiment of the invention provides a cable partial discharge signal identification method and device based on S transformation, which comprises the following steps: acquiring a partial discharge signal from a known source, and performing S transformation on the partial discharge signal to obtain a complex time-frequency matrix; performing module solving on the complex time-frequency matrix to obtain a module matrix, and performing singular value decomposition on the module matrix to obtain a singular value sequence of the module matrix; equally dividing the singular value sequence into at least two intervals according to the singular value sequence, calculating the ratio of the Shannon entropy of the singular value in each interval to the Shannon entropy of the singular value sequence, and establishing a partial discharge signal characteristic sample library by taking the ratio as a partial discharge signal characteristic vector; taking the partial discharge signal characteristic vectors in the partial discharge signal characteristic sample library as input, and constructing a multi-classification support vector machine model; taking the partial discharge signal as a sample, and training the support vector machine model to obtain a trained support vector machine model; inputting the characteristic vector of the partial discharge signal to be identified into a trained support vector machine model to obtain the source of the partial discharge signal to be identified, and obtaining a complex time-frequency matrix after S transformation is carried out on the partial discharge signal in the embodiment of the invention; performing module solving on the complex time-frequency matrix to obtain a module matrix, and performing singular value decomposition on the module matrix to obtain a singular value sequence of the module matrix; equally dividing the singular value sequence into at least two intervals according to the singular value sequence, calculating the ratio of the Shannon entropy of the singular value in each interval to the Shannon entropy of the singular value sequence, and establishing a partial discharge signal characteristic sample library by taking the ratio as a partial discharge signal characteristic vector; the method has the advantages that the method is simple in identification step, high in identification precision and high in identification speed, and solves the technical problems that the identification precision is greatly reduced and the required identification time is too long when the power cable line for transmitting the partial discharge signal is too long due to the fact that a waveform time difference method is usually adopted for positioning a power cable partial discharge signal source in the existing power industry.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without inventive exercise.
Fig. 1 is a schematic flowchart of an embodiment of a cable partial discharge signal identification method based on S transformation according to an embodiment of the present invention;
FIG. 2 is a waveform diagram of a partial discharge signal of a cable body according to an embodiment of the present invention;
fig. 3 is a waveform diagram of a partial discharge signal of a cable termination according to an embodiment of the present invention;
fig. 4 is a corona discharge signal of the switch cabinet provided by the embodiment of the invention;
fig. 5 is a waveform diagram of a surface discharge signal of the switch cabinet according to the embodiment of the invention;
fig. 6 is a schematic flowchart of another embodiment of a cable partial discharge signal identification method based on S transformation according to an embodiment of the present invention;
fig. 7 is a schematic structural diagram of a cable partial discharge signal identification device based on S transform according to an embodiment of the present invention.
Detailed Description
The embodiment of the invention provides a method and a device for identifying a cable partial discharge signal based on S transformation, which are used for solving the technical problems that the identification precision is greatly reduced and the required identification time is too long when a power cable line for transmitting the partial discharge signal is too long because a waveform time difference method is usually adopted for positioning a power cable partial discharge signal source in the power industry at present.
In order to make the objects, features and advantages of the present invention more obvious and understandable, the technical solutions in 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 obvious that the embodiments described below are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1, an embodiment of a cable partial discharge signal identification method based on S transform according to the present invention includes:
101. acquiring a partial discharge signal from a known source, and performing S transformation on the partial discharge signal to obtain a complex time-frequency matrix;
firstly, acquiring partial discharge signals of a cable, wherein partial discharge signals of different sources can be acquired through corresponding tests, and the partial discharge signals of different sources comprise a cable body partial discharge signal, a cable terminal partial discharge signal, a switch cabinet corona discharge signal, a switch cabinet surface discharge signal and the like, and are shown in fig. 2 as a cable body partial discharge signal waveform diagram; as shown in fig. 3, it is a waveform diagram of a partial discharge signal of a cable terminal; as shown in fig. 4, it is a corona discharge signal of the switch cabinet; fig. 5 shows a waveform diagram of a surface discharge signal of the switch cabinet. And carrying out S transformation on the obtained partial discharge signal to obtain a corresponding complex time-frequency matrix.
102. Performing module solving on the complex time-frequency matrix to obtain a module matrix, and performing singular value decomposition on the module matrix to obtain a singular value sequence of the module matrix;
after acquiring the partial discharge signal from a known source and performing S transformation on the partial discharge signal to obtain a complex time-frequency matrix, performing module solving on the complex time-frequency matrix to obtain a module matrix, and performing singular value decomposition on the module matrix to obtain a singular value sequence of the module matrix.
103. Equally dividing the singular value sequence into at least two intervals according to the singular value sequence, calculating the ratio of the Shannon entropy of the singular value in each interval to the Shannon entropy of the singular value sequence, and establishing a partial discharge signal characteristic sample library by taking the ratio as a partial discharge signal characteristic vector;
after a complex time-frequency matrix is subjected to modulus calculation to obtain a modulus matrix, singular value decomposition is carried out on the modulus matrix to obtain a singular value sequence of the modulus matrix, the singular value sequence is equally divided into at least two intervals according to the singular value sequence, for example, the singular value sequence is equally divided into 10 intervals, the ratio of the Shannon entropy of the singular value in each interval to the Shannon entropy of the whole singular value sequence is calculated, and the ratio is used as a partial discharge signal feature vector to establish a partial discharge signal feature sample library.
It should be noted that, when the partial discharge signal feature sample library is established, the source type of the partial discharge signal feature vector may be marked by using a binary number, that is, different sources of the partial discharge signal are marked, for example, the cable partial discharge signal, the cable terminal partial discharge signal, the corona discharge signal in the switch cabinet, and the surface discharge signal in the switch cabinet may be respectively marked as: (+1,+1),(+1, -1),(-1,+1),(-1, -1).
104. Taking the partial discharge signal characteristic vectors in the partial discharge signal characteristic sample library as input, and constructing a multi-classification support vector machine model;
equally dividing the singular value sequence into at least two intervals according to the singular value sequence, calculating the ratio of the Shannon entropy of the singular value in each interval to the Shannon entropy of the singular value sequence, establishing a partial discharge signal characteristic sample library by taking the ratio as a partial discharge signal characteristic vector, and then establishing a multi-classification support vector machine model by taking the partial discharge signal characteristic vector in the partial discharge signal characteristic sample library as input.
105. Taking the partial discharge signal as a sample, and training the support vector machine model to obtain a trained support vector machine model;
after the multi-classification support vector machine model is constructed, the trained support vector machine model can be obtained by acquiring partial discharge signals from different sources, inputting the acquired partial discharge signals from different sources into the support vector machine model as samples and training.
106. And inputting the characteristic vector of the partial discharge signal to be identified into a trained support vector machine model to obtain the source of the partial discharge signal to be identified.
And finally, when the source of the partial discharge signal of the cable needs to be identified, calculating the partial discharge signal to be identified to obtain a feature vector of the partial discharge signal, inputting the feature vector of the partial discharge signal into a trained support vector machine model, and obtaining the source of the partial discharge signal to be identified according to an output value of the support vector machine model.
In order to describe an embodiment of the cable partial discharge signal identification method based on the S transformation in detail, another embodiment of the cable partial discharge signal identification method based on the S transformation according to the embodiment of the present invention will be described in detail.
Referring to fig. 6, another embodiment of a cable partial discharge signal identification method based on S transform according to an embodiment of the present invention includes:
201. acquiring partial discharge signals of known sources, and performing S transformation on the partial discharge signals through a preset formula to obtain a complex time-frequency matrix;
firstly, acquiring partial discharge signals of a cable, wherein partial discharge signals of different sources can be acquired through corresponding tests, and the partial discharge signals of different sources comprise a cable body partial discharge signal, a cable terminal partial discharge signal, a switch cabinet corona discharge signal, a switch cabinet surface discharge signal and the like, and are shown in fig. 2 as a cable body partial discharge signal waveform diagram; as shown in fig. 3, it is a waveform diagram of a partial discharge signal of a cable terminal; as shown in fig. 4, it is a corona discharge signal of the switch cabinet; fig. 5 shows a waveform diagram of a surface discharge signal of the switch cabinet. S conversion is carried out on the obtained partial discharge signals through a preset formula I, and a corresponding complex time-frequency matrix can be obtained, wherein the preset formula I specifically comprises the following steps:
where H (kt) is a discrete time sequence of the partial discharge signal, ST is a complex time-frequency matrix obtained by transforming the discrete time sequence of the partial discharge signal by S, T is a sampling period of the discrete time sequence, N is a length of the discrete time sequence, H is a fourier transform of the discrete time sequence, j is an imaginary unit, k, N, m is 0,1, …, N-1.
202. Performing modulus calculation on the complex time-frequency matrix ST to obtain a modulus matrix STA, and performing singular value decomposition on the modulus matrix STA through a preset second formula to obtain a singular value sequence of the modulus matrix;
after acquiring partial discharge signals from a known source, performing S conversion on the partial discharge signals through a preset formula to obtain a complex time-frequency matrix, performing modulus on the complex time-frequency matrix ST to obtain a modulus matrix STA, and performing singular value decomposition on the modulus matrix STA through a preset second formula to obtain a singular value sequence of the modulus matrix, wherein the preset second formula specifically comprises the following steps:
where U and V are both N × order N orthogonal matrices, and D ═ diag (σ)1,σ2,…,σN) Is a diagonal matrix whose diagonal elements (σ)1,σ2,…,σN) Is the singular value of the matrix STA, uiAnd viThe vectors of the singular values of the ith column of the matrices U and V, respectively.
203. Equally dividing the singular value sequence into at least two intervals according to the singular value sequence, calculating the ratio of the Shannon entropy of the singular value in each interval to the Shannon entropy of the singular value sequence, and establishing a partial discharge signal characteristic sample library by using the ratio as a partial discharge signal characteristic vector through a preset third formula;
after a complex time-frequency matrix ST is subjected to modulus calculation to obtain a modulus matrix STA, singular value decomposition is carried out on the modulus matrix STA through a preset second formula to obtain a singular value sequence of the modulus matrix, the singular value sequence is equally divided into at least two intervals according to the singular value sequence, the ratio of the Shannon entropy of the singular value in each interval to the Shannon entropy of the singular value sequence is calculated, a local discharge signal characteristic sample library is established by using the ratio as a local discharge signal characteristic vector through a preset third formula, and the preset third formula specifically comprises the following steps:
λ＝[E1/E,E2/E,Eq/E…,EQ/E]；
It should be noted that, when the partial discharge signal feature sample library is established, the source type of the partial discharge signal feature vector may be marked by using a binary number, that is, different sources of the partial discharge signal are marked, for example, the cable partial discharge signal, the cable terminal partial discharge signal, the corona discharge signal in the switch cabinet, and the surface discharge signal in the switch cabinet may be respectively marked as: (+1,+1),(+1, -1),(-1,+1),(-1, -1).
204. Combining all source categories in the source categories of the partial discharge signal characteristic vectors in the partial discharge signal characteristic sample library by using a binary classification algorithm to form a plurality of sub-classifiers, and constructing a multi-classification support vector machine model by taking the partial discharge signal characteristic vectors in the partial discharge signal characteristic sample library as input;
equally dividing the singular value sequence into at least two intervals according to the singular value sequence, calculating the ratio of the Shannon entropy of the singular value in each interval to the Shannon entropy of the singular value sequence, establishing a partial discharge signal characteristic sample library by using the ratio as a partial discharge signal characteristic vector through a preset third formula, combining various source categories in the source categories of the partial discharge signal characteristic vector in the partial discharge signal characteristic sample library by using a classification algorithm to form a plurality of sub-classifiers, and establishing a multi-classification support vector machine model by using the partial discharge signal characteristic vector in the partial discharge signal characteristic sample library as input. Since the general support vector machine model is a two-classification support vector machine model and the source types of the partial discharge signals to be identified are four, it is necessary to combine each source type in the source types of the partial discharge signal feature vectors in the partial discharge signal feature sample library by using a two-classification algorithm to form 2 sub-classifiers SVM1 and SVM2, so as to expand the two-classification of the support vector machine to multiple classifications.
If 4 kinds of discharge signals, namely a cable body partial discharge signal, a cable terminal partial discharge signal, a corona discharge signal in a switch cabinet and a surface discharge signal in the switch cabinet, are respectively represented by A, B, C and D, the corresponding relation between the output result of the SVM1 and the type of the discharge signal of the SVM2 is shown in Table 1.
TABLE 1
205. Randomly selecting partial discharge signals of different sources with the same quantity as training samples according to different sources of the partial discharge signals, and inputting the training samples into a support vector machine model for training to obtain a trained support vector machine model;
after the multi-classification support vector machine model is constructed, the partial discharge signals of the same number and different sources are randomly selected as training samples according to different sources of the partial discharge signals, namely, the partial discharge signals of the same number are randomly selected from four partial discharge signals, such as 50 cable body partial discharge signals, 50 cable terminal partial discharge signals, 50 switch cabinet corona discharge signals and 50 switch cabinet surface discharge signals, and the training samples are input into the support vector machine model for training to obtain the trained support vector machine model.
206. And inputting the characteristic vector of the partial discharge signal to be identified into a trained support vector machine model to obtain an output value, and comparing the output value with the binary number label of the source category of the characteristic vector of the partial discharge signal to obtain the source of the partial discharge signal to be identified.
Finally, when the source of the partial discharge signal of the cable needs to be identified, the partial discharge signal to be identified is calculated to obtain a feature vector of the partial discharge signal, the feature vector of the partial discharge signal to be identified is input into a trained support vector machine model to obtain an output value, the output value is compared with a binary number label of the source category of the feature vector of the partial discharge signal to obtain the source of the partial discharge signal to be identified, namely as shown in table 1, the output values are (+1, +1), (+1, -1), (-1, +1), (-1, -1) respectively correspond to the cable partial discharge signal, the cable terminal partial discharge signal, the corona discharge signal in the switch cabinet, and the surface discharge signal in the switch cabinet.
In order to describe the cable partial discharge signal identification method based on the S-transform in detail, a cable partial discharge signal identification device based on the S-transform according to an embodiment of the present invention will be described in detail.
Referring to fig. 7, an apparatus for identifying a partial discharge signal of a cable based on S transform according to an embodiment of the present invention includes:
the transformation module 301 is configured to obtain a local discharge signal from a known source, and perform S transformation on the local discharge signal to obtain a complex time-frequency matrix;
the decomposition module 302 is configured to perform modulo calculation on the complex time-frequency matrix to obtain a modulus matrix, and perform singular value decomposition on the modulus matrix to obtain a singular value sequence of the modulus matrix;
the calculation module 303 is configured to equally divide the singular value sequence into at least two intervals according to the singular value sequence, calculate a ratio of a Shannon entropy of the singular value in each interval to a Shannon entropy of the singular value sequence, and establish a partial discharge signal feature sample library by using the ratio as a partial discharge signal feature vector;
a building module 304, configured to build a multi-classification support vector machine model by using the partial discharge signal feature vector in the partial discharge signal feature sample library as an input;
the training module 305 is configured to train the support vector machine model by using the partial discharge signal as a sample, so as to obtain a trained support vector machine model;
the input module 306 is configured to input the feature vector of the partial discharge signal to be identified into the trained support vector machine model, so as to obtain a source of the partial discharge signal to be identified.
It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the above-described systems, apparatuses and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
In the several embodiments provided in the present application, it should be understood that the disclosed system, apparatus and method may be implemented in other manners. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the units is only one logical division, and other divisions may be realized in practice, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.
The units described as separate parts may or may not be physically separate, and parts displayed 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 can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.
The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
The above-mentioned embodiments are only used for illustrating the technical solutions of the present invention, and not for limiting the same; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. A cable partial discharge signal identification method based on S transformation is characterized by comprising the following steps:
acquiring partial discharge signals from known sources, and performing S transformation on the partial discharge signals to obtain a complex time-frequency matrix;
performing module solving on the complex time-frequency matrix to obtain a module matrix, and performing singular value decomposition on the module matrix to obtain a singular value sequence of the module matrix;
equally dividing the singular value sequence into at least two intervals according to the singular value sequence, calculating the ratio of the Shannon entropy of the singular value in each interval to the Shannon entropy of the singular value sequence, and establishing a partial discharge signal characteristic sample library by taking the ratio as a partial discharge signal characteristic vector;
taking the partial discharge signal characteristic vectors in the partial discharge signal characteristic sample library as input, and constructing a multi-classification support vector machine model;
taking the partial discharge signal as a sample, and training the support vector machine model to obtain a trained support vector machine model;
and inputting the characteristic vector of the partial discharge signal to be identified into the trained support vector machine model to obtain the source of the partial discharge signal to be identified.
2. The S-transform-based cable partial discharge signal identification method according to claim 1, wherein the partial discharge signals include a cable body partial discharge signal, a cable terminal partial discharge signal, a switch cabinet corona discharge signal, and a switch cabinet surface discharge signal.
3. The method for identifying cable partial discharge signals based on S-transform as claimed in claim 1, wherein the obtaining partial discharge signals of known source and S-transform the partial discharge signals to obtain a complex time-frequency matrix comprises:
obtaining a partial discharge signal from a known source, and performing S transformation on the partial discharge signal by using a preset formula to obtain a complex time-frequency matrix, wherein the preset formula I specifically comprises:
where H (kt) is a discrete time sequence of the partial discharge signal, ST is a complex time-frequency matrix obtained by transforming the discrete time sequence of the partial discharge signal by S, T is a sampling period of the discrete time sequence, N is a length of the discrete time sequence, H is a fourier transform of the discrete time sequence, j is an imaginary unit, k, N, m is 0,1, …, N-1.
4. The S-transform-based cable partial discharge signal identification method of claim 3, wherein the performing a modulo operation on the complex time-frequency matrix to obtain a module matrix, and performing a singular value decomposition on the module matrix to obtain a singular value sequence of the module matrix comprises:
performing module solving on the complex time-frequency matrix ST to obtain a module matrix STA, and performing singular value decomposition on the module matrix STA through a preset second formula to obtain a singular value sequence of the module matrix, wherein the preset second formula specifically comprises:
where U and V are both N × order N orthogonal matrices, and D ═ diag (σ)1,σ2,…,σN) Is a diagonal matrix whose diagonal elements (σ)1,σ2,…,σN) Is the singular value of the matrix STA, uiAnd viThe vectors of the singular values of the ith column of the matrices U and V, respectively.
5. The S-transform-based cable partial discharge signal identification method according to claim 4, wherein the equally dividing the singular value sequence into at least two intervals according to the singular value sequence, calculating a ratio of Shannon entropy of singular values in each interval to Shannon entropy of the singular value sequence, and establishing a partial discharge signal feature sample library using the ratio as a partial discharge signal feature vector comprises:
equally dividing the singular value sequence into at least two intervals according to the singular value sequence, calculating a ratio of the Shannon entropy of the singular value in each interval to the Shannon entropy of the singular value sequence, and establishing a partial discharge signal characteristic sample library by taking the ratio as a partial discharge signal characteristic vector through a preset third formula, wherein the preset third formula specifically comprises the following steps:
λ＝[E1/E,E2/E,Eq/E…,EQ/E]；
6. The S-transform-based cable partial discharge signal identification method of claim 5, wherein the source category of the partial discharge signal feature vector is marked by a two-bit binary number.
7. The S-transform-based cable partial discharge signal identification method according to claim 6, wherein the constructing a multi-class support vector machine model with the partial discharge signal feature vectors in the partial discharge signal feature sample library as inputs comprises:
and combining all the source categories of the partial discharge signal characteristic vectors in the partial discharge signal characteristic sample library by using a binary classification algorithm to form a plurality of sub-classifiers, and constructing a multi-classification support vector machine model by taking the partial discharge signal characteristic vectors in the partial discharge signal characteristic sample library as input.
8. The method for identifying the cable partial discharge signal based on the S transformation according to claim 7, wherein the training the support vector machine model by using the partial discharge signal as a sample to obtain the trained support vector machine model comprises:
and randomly selecting the partial discharge signals of the same number and different sources as training samples according to different sources of the partial discharge signals, and inputting the training samples into the support vector machine model for training to obtain the trained support vector machine model.
9. The S-transform-based cable partial discharge signal identification method according to claim 8, wherein the inputting the feature vector of the partial discharge signal to be identified into the trained support vector machine model to obtain the source of the partial discharge signal to be identified comprises:
and inputting the characteristic vector of the partial discharge signal to be identified into the trained support vector machine model to obtain an output value, and comparing the output value with the binary number label of the source category of the characteristic vector of the partial discharge signal to obtain the source of the partial discharge signal to be identified.
10. An S-transform-based cable partial discharge signal identification device, comprising:
the transformation module is used for acquiring partial discharge signals from known sources and carrying out S transformation on the partial discharge signals to obtain a complex time-frequency matrix;
the decomposition module is used for performing module solving on the complex time-frequency matrix to obtain a module matrix and performing singular value decomposition on the module matrix to obtain a singular value sequence of the module matrix;
the calculation module is used for equally dividing the singular value sequence into at least two intervals according to the singular value sequence, calculating the ratio of the Shannon entropy of the singular value in each interval to the Shannon entropy of the singular value sequence, and establishing a partial discharge signal characteristic sample library by taking the ratio as a partial discharge signal characteristic vector;
the construction module is used for taking the partial discharge signal characteristic vector in the partial discharge signal characteristic sample library as input to construct a multi-classification support vector machine model;
the training module is used for training the support vector machine model by taking the partial discharge signal as a sample to obtain a trained support vector machine model;
and the input module is used for inputting the characteristic vector of the partial discharge signal to be identified into the trained support vector machine model to obtain the source of the partial discharge signal to be identified.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710035996.3A CN106778692B (en) | 2017-01-17 | 2017-01-17 | Cable partial discharge signal identification method and device based on S transformation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710035996.3A CN106778692B (en) | 2017-01-17 | 2017-01-17 | Cable partial discharge signal identification method and device based on S transformation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106778692A CN106778692A (en) | 2017-05-31 |
| CN106778692B true CN106778692B (en) | 2020-07-03 |
Family
ID=58944005
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710035996.3A Active CN106778692B (en) | 2017-01-17 | 2017-01-17 | Cable partial discharge signal identification method and device based on S transformation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106778692B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107255779A (en) * | 2017-06-15 | 2017-10-17 | 国网河南省电力公司济源供电公司 | A kind of high-tension switch cabinet Analysis of Partial Discharge method |
| CN107862320A (en) * | 2017-11-28 | 2018-03-30 | 广东电网有限责任公司珠海供电局 | A kind of porcelain shell for cable terminal Infrared Image Features vector extracting method |
| CN114325270B (en) * | 2021-12-30 | 2023-07-04 | 广东电网有限责任公司江门供电局 | Switch cabinet partial discharge monitoring method and device |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103076547A (en) * | 2013-01-24 | 2013-05-01 | 安徽省电力公司亳州供电公司 | Method for identifying GIS (Gas Insulated Switchgear) local discharge fault type mode based on support vector machines |
| CN103077402A (en) * | 2012-12-28 | 2013-05-01 | 湖北省电力公司电力科学研究院 | Transformer partial-discharging mode recognition method based on singular value decomposition algorithm |
| CN103091612A (en) * | 2013-01-16 | 2013-05-08 | 重庆市电力公司电力科学研究院 | Separation and recognition algorithm for transformer oiled paper insulation multiple partial discharging source signals |
| CN106250904A (en) * | 2016-05-18 | 2016-12-21 | 国网新疆电力公司电力科学研究院 | Based on Power Disturbance analyser and the sorting technique of improving S-transformation |
-
2017
- 2017-01-17 CN CN201710035996.3A patent/CN106778692B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103077402A (en) * | 2012-12-28 | 2013-05-01 | 湖北省电力公司电力科学研究院 | Transformer partial-discharging mode recognition method based on singular value decomposition algorithm |
| CN103091612A (en) * | 2013-01-16 | 2013-05-08 | 重庆市电力公司电力科学研究院 | Separation and recognition algorithm for transformer oiled paper insulation multiple partial discharging source signals |
| CN103076547A (en) * | 2013-01-24 | 2013-05-01 | 安徽省电力公司亳州供电公司 | Method for identifying GIS (Gas Insulated Switchgear) local discharge fault type mode based on support vector machines |
| CN106250904A (en) * | 2016-05-18 | 2016-12-21 | 国网新疆电力公司电力科学研究院 | Based on Power Disturbance analyser and the sorting technique of improving S-transformation |
Non-Patent Citations (4)
| Title |
|---|
| 《Feature extraction of GIS partial discharge signal based on S-transform and singular value decomposition》;Dangdang Dai等;《IET Digital Library》;20160818;摘要，第186页右栏最后一段，第4.1节，第5.1节第2段 * |
| 《S变换在电力电缆局部放电信号时频分析中的应用》;郭灿新等;《电工技术学报》;20101130;第25卷(第11期);第11页左栏第1-3段 * |
| 《频率切片小波变换在局部放电信号分析中的应用》;张宇辉等;《高电压技术》;20150731;第41卷(第7期);第2283-2293页 * |
| Dangdang Dai等.《Feature extraction of GIS partial discharge signal based on S-transform and singular value decomposition》.《IET Digital Library》.2016,摘要，第186页右栏最后一段，第4.1节，第5.1节第2段. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106778692A (en) | 2017-05-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Karimzadeh et al. | Method for determining utility and consumer harmonic contributions based on complex independent component analysis | |
| Singh et al. | Application of fractional Fourier transform for classification of power quality disturbances | |
| CN103076547B (en) | Method for identifying GIS (Gas Insulated Switchgear) local discharge fault type mode based on support vector machines | |
| Song et al. | Multiple event detection and recognition for large-scale power systems through cluster-based sparse coding | |
| Chen et al. | An improved Hurst parameter estimator based on fractional Fourier transform | |
| CN106778692B (en) | Cable partial discharge signal identification method and device based on S transformation | |
| CN106443380B (en) | A kind of distribution cable local discharge signal recognition methods and device | |
| CN103149514B (en) | Method for identifying multi-insulated defect mode in GIS (gas insulated switchgear) | |
| Ren et al. | Power system event classification and localization using a convolutional neural network | |
| CN111080105A (en) | Transformer area user-to-user relationship identification method and system based on voltage time sequence data | |
| CN106203103B (en) | File virus detection method and device | |
| CN105701470A (en) | Analog circuit fault characteristic extraction method based on optimal wavelet packet decomposition | |
| Saber et al. | Discrete wavelet transform and support vector machine‐based parallel transmission line faults classification | |
| CN102184451B (en) | Multi-wavelet fusion feature based neuron action feature extraction method | |
| CN110610226A (en) | Generator fault prediction method and device | |
| CN116756594A (en) | Method, system, equipment and medium for detecting abnormal points of power grid data | |
| Cui et al. | Spatio-temporal synchrophasor data characterization for mitigating false data injection in smart grids | |
| CN114118149A (en) | Induction motor fault diagnosis system based on finite element simulation and symmetric feature migration | |
| Petladwala et al. | Canonical correlation based feature extraction with application to anomaly detection in electric appliances | |
| Ye et al. | Distinguishing chaotic time series from noise: A random matrix approach | |
| CN116298509A (en) | Online identification method for harmonic resonance of power system | |
| CN115293303A (en) | High-voltage transmission line network monitoring method, system, equipment and medium | |
| Hsieh et al. | Identification of partial discharge location using probabilistic neural networks and the fuzzy c-means clustering approach | |
| Sima et al. | A framework for automatically cleansing overvoltage data measured from transmission and distribution systems | |
| Seth et al. | A test of granger non-causality based on nonparametric conditional independence |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |