CN116482526A - Analysis system for non-fault phase impedance relay - Google Patents
Analysis system for non-fault phase impedance relay Download PDFInfo
- Publication number
- CN116482526A CN116482526A CN202310477568.1A CN202310477568A CN116482526A CN 116482526 A CN116482526 A CN 116482526A CN 202310477568 A CN202310477568 A CN 202310477568A CN 116482526 A CN116482526 A CN 116482526A
- Authority
- CN
- China
- Prior art keywords
- signal
- feature
- model
- signals
- data
- 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.)
- Pending
Links
- 238000004458 analytical method Methods 0.000 title claims abstract description 35
- 238000012545 processing Methods 0.000 claims abstract description 51
- 238000000034 method Methods 0.000 claims abstract description 37
- 230000008569 process Effects 0.000 claims abstract description 20
- 238000012549 training Methods 0.000 claims description 24
- 238000005070 sampling Methods 0.000 claims description 22
- 238000001914 filtration Methods 0.000 claims description 20
- 239000013598 vector Substances 0.000 claims description 17
- 238000012360 testing method Methods 0.000 claims description 15
- 238000004364 calculation method Methods 0.000 claims description 13
- 238000000605 extraction Methods 0.000 claims description 12
- 238000010606 normalization Methods 0.000 claims description 10
- 238000007619 statistical method Methods 0.000 claims description 9
- 230000009466 transformation Effects 0.000 claims description 9
- 230000006870 function Effects 0.000 claims description 7
- 239000011159 matrix material Substances 0.000 claims description 5
- 238000001228 spectrum Methods 0.000 claims description 4
- 230000008859 change Effects 0.000 claims description 3
- 238000005316 response function Methods 0.000 claims description 3
- 238000012706 support-vector machine Methods 0.000 description 6
- 238000004422 calculation algorithm Methods 0.000 description 3
- 238000003745 diagnosis Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000003595 spectral effect Effects 0.000 description 3
- 230000005856 abnormality Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 238000007476 Maximum Likelihood Methods 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000010219 correlation analysis Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000007477 logistic regression Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000010187 selection method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/327—Testing of circuit interrupters, switches or circuit-breakers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/10—Pre-processing; Data cleansing
-
- 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
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
- G06F2218/04—Denoising
- G06F2218/06—Denoising by applying a scale-space analysis, e.g. using wavelet analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
-
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
- Y04S10/52—Outage or fault management, e.g. fault detection or location
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Emergency Protection Circuit Devices (AREA)
Abstract
The invention discloses an analysis system for a non-fault phase impedance relay, which comprises a data acquisition module, a signal processing module and a decision module, wherein the data acquisition module is connected with the signal processing module, the signal processing module is connected with the decision module, the data acquisition module acquires input and output signals of the relay, the signal processing module processes and analyzes the signals acquired by the data acquisition module, and the decision module is used for processing data based on the signal processing module.
Description
Technical Field
The invention relates to the technical field of relays, in particular to an analysis system for a non-fault phase impedance relay.
Background
Phase-impedance relays are a common protection device for detecting faults and protecting in electrical power systems. However, in actual operation, the phase-resistance relay may also have phase-resistance abnormality in a non-failure state for various reasons. Such anomalies may lead to false positives or even damage to equipment.
Therefore, an analysis system is needed to determine whether the phase resistance of the relay is normal or not, and a technical solution is provided.
Disclosure of Invention
In order to overcome the above-mentioned drawbacks of the prior art, an embodiment of the present invention provides an analysis system for a non-fault phase impedance relay, which can accurately determine whether the relay has a phase impedance anomaly by combining a data acquisition module, a signal processing module and a decision module, so as to solve the problems set forth in the background art.
In order to achieve the above purpose, the present invention provides the following technical solutions:
an analysis system for a non-fault phase impedance relay comprises a data acquisition module, a signal processing module and a decision module, wherein the data acquisition module is connected with the signal processing module, and the signal processing module is connected with the decision module; the data acquisition module acquires input and output signals of the relay, converts the signals into acceptable electric signals, and is connected into the data acquisition card, and sequentially performs signal filtering, denoising and phase resistance calculation on the acquired signals;
the signal processing module processes and analyzes the signals acquired by the data acquisition module, and a model is built for diagnosing faults through signal analysis and feature extraction;
the decision module is used for predicting and deciding the risk of the working state and faults of the relay based on the data processed by the signal processing module.
As a further technical scheme of the invention, the specific steps of denoising processing in the data acquisition module are as follows:
step A1, frequency domain filtering: performing fast Fourier transform on the acquired data, converting the signal into a frequency domain, and filtering the frequency domain signal by adopting a low-pass, high-pass or band-pass filter according to the signal-to-noise ratio, wherein the formula is X (f) =H (f) Y (f);
wherein X (f) represents the filtered frequency domain signal, H (f) is the frequency response function of the filter, and (f) is the frequency domain representation of the original signal;
step A2, wavelet transformation: wavelet transformation is carried out on the acquired data, signals are decomposed into sub-bands with different scales, noise signals are removed by utilizing the characteristics of the different sub-bands, and the formula is that
Wherein x (n) is an original signal, h (k) and g (k) are wavelet coefficients, and y (n) is a denoised signal;
step A3, statistical method: statistical analysis is carried out on the collected data, and denoising processing is carried out on the data through the statistical parameters of mean value, variance, skewness and kurtosis, wherein the formula is that
Wherein,,for the mean value, n is the data length, X i Is the ith data.
As a further technical scheme of the invention, the specific steps of phase resistance calculation are as follows:
step B1, collecting current and voltage signals: collecting current and voltage signals in a power system through a current transformer and a voltage transformer, and obtaining accurate current and voltage waveform data through signal filtering and denoising;
step B2, phase synchronization: the current and voltage signals are subjected to phase synchronization, so that the sampling moments of the current and voltage signals are completely consistent;
step B3, fourier transformation: performing fast Fourier transform on the current and voltage signals, and converting the signals into a frequency domain, wherein the formula is
Wherein X (f) represents a frequency domain signal, N is the number of sampling points, N is the serial number of the sampling points, X (N) is an original signal, f is the frequency, and T is the sampling time interval;
step B4, calculating phase impedance: according to the frequency domain signals of the current and the voltage, the phase resistance in the power system is obtained by adopting the Ohm law and the kirchhoff law for calculation, and the formula is that
Wherein Z (f) is phase resistance, U (f) is a frequency domain signal of voltage, and I (f) is a frequency domain signal of current.
As a further technical scheme of the invention, the specific method for extracting the characteristics in the signal processing module comprises the following steps: before the feature extraction, the original signal needs to be divided into a time window with a certain length, and then the statistical features of the signal are extracted in the time domain, the frequency spectrum information of the signal is extracted in the frequency domain, and the change information of the signal is extracted in the wavelet domain in sequence.
As a further technical scheme of the invention, the specific process of establishing the model in the signal processing module is as follows: selecting a proper model type from the characteristic data set selected by the characteristic selection module; setting parameters of the selected model; training the selected model by using a training data set, testing the trained model by using a testing data set, evaluating the performance of the model and optimizing the model, and storing the optimized model into a model library for subsequent use.
As a further technical scheme of the invention, the decision module is realized by adopting a computer or an embedded processor device, and the specific method of the decision module is as follows:
selecting the characteristics obtained by processing by the signal processing module;
based on the feature selection, establishing a corresponding model based on the existing data;
on the basis of establishing a model, corresponding decision rules are determined, and corresponding decisions are carried out according to the result of model prediction.
As a further technical scheme of the invention, the specific steps of establishing the model are as follows:
step C1, preparing a data set, dividing the data set into a training set and a testing set, wherein the training set is used for training a model, and the testing set is used for evaluating the performance of the model;
step C2, adopting a formulaNormalizing the features, wherein x represents the original value of a feature, x min And x max Respectively represent the minimum and maximum values of the feature, x norm Representing the normalized eigenvalue;
step C3, after feature normalization, model training is carried out;
and step C4, after model training is completed, model prediction is performed.
The invention provides a technical effect and advantages of an analysis system for a non-fault phase impedance relay, which are as follows: the invention adopts automatic data acquisition and signal processing technology, does not need manual intervention, and has simple and convenient operation; the invention adopts common hardware equipment and software algorithm, has lower cost and is easy to popularize and apply, and the invention can accurately judge whether the phase resistance abnormality exists in the relay, thereby avoiding the occurrence of misjudgment and equipment damage.
Drawings
Fig. 1 is a schematic structural diagram of an analysis system for a non-fault phase-impedance relay according to the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. 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.
An analysis system for a non-fault phase impedance relay comprises a data acquisition module, a signal processing module and a decision module, wherein the data acquisition module is connected with the signal processing module, and the signal processing module is connected with the decision module;
the data acquisition module acquires input and output signals of the relay, converts the signals into acceptable electric signals, and is connected into the data acquisition card, and sequentially performs signal filtering, denoising and phase resistance calculation on the acquired signals;
the data acquisition module is a device for acquiring input and output signals of the relay, and in the implementation, a data acquisition card of NI company, such as NIUSB-6211 data acquisition card, is adopted. The data acquisition card supports a variety of input and output signal types including analog signals, digital signals, frequency and count, and the like. For input and output signals of the relay, the signals can be converted into acceptable electric signals by adopting accessories such as a sensor and an interface board provided by NI company, and the signals are connected into a data acquisition card.
In practice, the input and output signals of the relay are often subject to various disturbances, such as power supply noise, sensor drift, etc. To reduce the effect of these disturbances on the signal, we use digital filtering techniques to filter the signal. Among them, the usual digital filters are low-pass filters, band-stop filters, and the like.
In addition to signal interference, the input and output signals of the relay may also be affected by noise. To reduce the effect of noise on the signal, we use denoising techniques to process the signal. Among them, the common denoising techniques include wavelet denoising, adaptive filtering, mean filtering, and the like. For the input and output signals of the relay, the proper denoising technology and parameters can be selected for processing according to specific situations.
After the filtered and denoised signals are collected, we can calculate the phase resistance of the relay. The phase impedance refers to impedance characteristics of the relay during operation, including impedance magnitude, phase angle, and the like. For the input and output signals of the relay, we calculate its phase resistance using the following formula:
z=v/I; wherein Z is phase impedance, V is voltage of relay output signal, and I is current of relay input signal.
The signal filtering in the data acquisition module refers to filtering processing of the original signal acquired from the impedance relay so as to eliminate noise interference and improve signal quality, thereby improving the accuracy and reliability of subsequent analysis, and the implementation process is as follows:
pretreatment: firstly, preprocessing the collected original signals, including DC component removal, normalization processing and the like. The dc component is removed in order to avoid drift and offset of the signal on the dc component, affecting subsequent signal analysis. The normalization process is to eliminate the signal amplitude differences between the different acquisition points.
And (3) filtering: the signal is filtered using a digital filter. Common digital filters include low pass filters, high pass filters, band reject filters, and the like. And selecting proper filter types and filter parameters, and performing filtering processing according to the signal characteristics of the impedance relay.
And (3) performing inverse normalization treatment: and carrying out inverse normalization processing on the filtered signal to restore the signal to the range of the original signal amplitude. The specific formula is as follows:
removing the direct current component:
where x (T) is the original signal and T is the sampling time.
Normalization:
wherein x' (t) is normalizedThe processed signal, max (x), is the maximum value of the original signal.
Digital filter: digital filters are typically expressed in the form of differential equations,
wherein x [ N ] is an input signal, y [ N ] is an output signal, b [ i ] and a [ j ] are respectively a forward coefficient and a feedback coefficient of the filter, and M and N are respectively the orders of the forward coefficient and the feedback coefficient.
And (3) performing inverse normalization treatment:
x "(t) =x' (t) ×max (x), where x" (t) is the inverse normalized signal.
The denoising processing in the data acquisition module comprises the following specific steps:
step A1, frequency domain filtering: the acquired data is subjected to a Fast Fourier Transform (FFT) to convert the signal to the frequency domain. Filtering the frequency domain signal by adopting a low-pass, high-pass or band-pass filter according to the signal-to-noise ratio (SNR), wherein the formula is X (f) =H (f) Y (f);
wherein X (f) represents the filtered frequency domain signal, H (f) is the frequency response function of the filter, and (f) is the frequency domain representation of the original signal;
step A2, wavelet transformation: wavelet transformation is carried out on the acquired data, signals are decomposed into sub-bands with different scales, noise signals are removed by utilizing the characteristics of the different sub-bands, and the formula is that
Wherein x (n) is an original signal, h (k) and g (k) are wavelet coefficients, and y (n) is a denoised signal;
step A3, statistical method: statistical analysis is carried out on the collected data, and denoising processing is carried out on the data through statistical parameters such as mean value, variance, skewness, kurtosis and the like, wherein the formula is that
Wherein,,for the mean value, n is the data length, X i Is the ith data.
Phase resistance calculation is an important component of the data acquisition module, and aims to calculate phase resistance in the power system for subsequent fault diagnosis and positioning. The following is a specific process and formula for implementation:
step B1, collecting current and voltage signals: collecting current and voltage signals in a power system through a current transformer and a voltage transformer, and obtaining accurate current and voltage waveform data through signal filtering and denoising;
step B2, phase synchronization: the current and voltage signals are subjected to phase synchronization, so that the sampling moments of the current and voltage signals are ensured to be completely consistent, and phase errors are avoided in the calculation process;
step B3, fourier transformation: performing fast Fourier transform on the current and voltage signals, and converting the signals into a frequency domain, wherein the formula is
Wherein X (f) represents a frequency domain signal, N is the number of sampling points, N is the serial number of the sampling points, X (N) is an original signal, f is the frequency, and T is the sampling time interval;
step B4, calculating phase impedance: according to the frequency domain signals of the current and the voltage, the famous Omm law and the kirchhoff law are adopted for calculation to obtain the phase resistance in the power system, and the formula is that
Wherein Z (f) is phase resistance, U (f) is a frequency domain signal of voltage, and I (f) is a frequency domain signal of current.
By the method, the phase resistance in the power system can be calculated, and an accurate data basis is provided for subsequent fault diagnosis and positioning.
The signal processing module processes and analyzes the signals acquired by the data acquisition module, and a model is built for diagnosing faults through signal analysis and feature extraction; in the implementation, the signal processing module is implemented by a computer or an embedded processor. For the acquired signals, we can process by the following specific procedure:
1. signal analysis:
first, we need to analyze the collected signals to understand the characteristics and regularity of the signals. The signal analysis adopts two methods of time domain analysis and frequency domain analysis. Time domain analysis may obtain information such as the waveform and amplitude of a signal by sampling and reconstructing the signal. The frequency domain analysis may then obtain information such as the frequency spectrum and frequency distribution of the signal by performing fourier transform on the signal. The purpose is to extract useful characteristic information for subsequent fault diagnosis and location by analyzing the collected current and voltage signals. The following is a specific process implemented:
collecting current and voltage signals: the current and voltage signals in the power system are collected through the current transformer and the voltage transformer, and accurate current and voltage waveform data are obtained through signal filtering and denoising processing.
Phase synchronization: and the current and voltage signals are subjected to phase synchronization, so that the sampling moments of the current and voltage signals are completely consistent, and phase errors in the analysis process are avoided.
Frequency characteristic extraction: useful frequency characteristic information is extracted through frequency domain analysis of the current and voltage signals. For example, the power spectral densities of the current and voltage signals may be calculated, and then characteristic information such as the valley, peak, and frequency distribution therein may be extracted.
The formula:
wherein P (f) is the power spectral density, X (f) is the frequency domain signal, and R is the resistance.
Extracting time domain features: extracting useful time domain characteristic information through time domain analysis of current and voltage signals;
the formula:wherein mu is the mean value, N is the sampling point number, N is the sampling point sequence number, and x (N) is the original signal.
And (3) extracting statistical characteristics: extracting useful statistical characteristic information through statistical analysis of current and voltage signals; the formula:wherein, gamma 1 The deviation is that mu is the mean value, N is the number of sampling points, N is the serial number of the sampling points, x (N) is the original signal, and sigma is the standard deviation.
2. Feature extraction
On the basis of signal analysis, the working state of the relay is analyzed by extracting the characteristics of the signals. By means of feature extraction, important features of the working state of the relay are extracted from the collected signals.
The process of feature extraction can be divided into the following steps:
determining a time window length of the signal: prior to feature extraction, the original signal needs to be divided into time windows of a certain length. The length of the time window needs to be determined according to the specific application.
Calculating time domain features: time domain features refer to extracting statistical features of signals in the time domain. Some common time domain features include mean, variance, standard deviation, maximum, minimum, peak-to-peak, etc.
Taking the mean value and standard deviation of the signals as an example, the calculation formula is as follows:
wherein x is i The value of the ith sampling point is given, and N is the number of sampling points in the window.
Calculating frequency domain features: the frequency domain feature refers to extracting spectrum information of a signal in a frequency domain. Some common frequency domain features include peak frequencies, band energy ratios, spectral entropy, etc.
Taking the peak frequency of the signal as an example, the calculation formula is as follows:
wherein X (f) i ) For a signal at a frequency f i Fourier transform values at T are the sampling periods of the signal.
Calculating wavelet domain features: wavelet domain features refer to extracting change information of signals in the wavelet domain. Some common wavelet domain features include the mean, standard deviation, energy, entropy, etc. of wavelet coefficients.
Taking the energy of the wavelet coefficient of the signal as an example, the calculation formula is as follows:
wherein c i Is the ith wavelet coefficient.
3. Modeling
On the basis of feature extraction, a corresponding model is established to analyze the working state and the characteristics of the relay, and the working state and the fault risk of the relay can be predicted more accurately by establishing the model.
The specific process of establishing the model in the signal processing module is as follows: selecting a proper model type from the characteristic data set selected by the characteristic selection module; setting parameters of the selected model; training the selected model by using a training data set, testing the trained model by using a testing data set, evaluating the performance of the model and optimizing the model, and storing the optimized model into a model library for subsequent use.
The decision module is used for predicting and deciding the risk of the working state and faults of the relay based on the data processed by the signal processing module. In the implementation, the decision module is implemented by a computer or an embedded processor. For the data processed by the signal processing module, the following specific processes can be used for prediction and decision:
the characteristics obtained by processing the signal processing module are selected, and common characteristic selection methods include correlation analysis, information gain and the like. Through feature selection, features which have important influences on the working state and fault risk of the relay can be selected from the features obtained through processing;
on the basis of feature selection, a corresponding model is established based on the existing data, and common modeling methods include logistic regression, naive Bayes, support vector machines and the like. By establishing a model, the working state and fault risk of the relay can be predicted;
on the basis of establishing a model, corresponding decision rules are determined, corresponding decisions are carried out according to the result of model prediction, and the common decision rules comprise a maximum likelihood decision rule, a minimum risk decision rule and the like. By means of the decision rule, corresponding decisions can be made according to the model prediction results.
The feature selection in the decision module is to screen and select the feature obtained by processing the signal processing module, wherein the feature selection in the information gain comprises the following steps:
step S1, calculating the information entropy of each feature, wherein the information entropy can measure the information content of one feature and is calculated by using the following formula:
wherein X represents a feature, p (X i ) Representing that the characteristic X takes the value X i Is a probability of (2).
Step S2, calculating a conditional entropy corresponding to each feature, wherein the conditional entropy can measure the influence degree of one feature on classification and is calculated by using the following formula:
wherein Y represents a class label, X represents a feature, p (X i ) Representing that the characteristic X takes the value X i Is H (y|x=x i ) Representing that the characteristic X takes the value X i The conditional entropy of the class label Y.
Step S3, calculating the information gain of each feature, wherein the information gain can measure the contribution degree of one feature to classification and is calculated by using the following formula:
IG(Y，X)＝H(Y)-H(Y|X)；
wherein Y represents a classification label, X represents a feature, H (Y) represents information entropy of the classification label Y, and H (Y|X) represents conditional entropy corresponding to the feature X;
in step S4, a feature having a large information gain is selected, and a feature having a large information gain is selected as a final feature, and a threshold is usually set to select a feature, for example, only a feature having an information gain of 0.1 or more is selected.
The modeling in the decision module is to implement prediction and decision making for the non-fault phase impedance relay, and machine learning algorithms can be used to build the model. In implementations, the model may be built using a variety of machine learning algorithms, including decision trees, support vector machines, neural networks, and the like. Taking a support vector machine as an example, the specific steps of building a model are as follows:
step C1, preparing a data set, dividing the data set into a training set and a testing set, and using 80% of data as the training set and 20% of data as the testing set. The training set is used for training the model, and the testing set is used for evaluating the performance of the model;
step C2, before modeling using a support vector machine, the features need to be normalized for better training. Normalizing the features by using a minimum-maximum normalization method, and adopting a formulaWherein x represents the original value of a feature, x min And x max Respectively represent the minimum and maximum values of the feature, x norm Representation normalizedIs a characteristic value of (2);
and step C3, after feature normalization, model training can be performed by using a support vector machine. The support vector machine is a two-class model that can divide samples into positive and negative classes. In model training, appropriate kernel functions and hyper-parameters need to be selected. Using a radial basis function as a kernel function, and selecting an optimal super parameter by using a grid search method;
and step C4, after model training is completed, model prediction can be performed by using the test set. For each test sample, its characteristic value is input into the model, which gives a probability value that the sample belongs to the positive or negative class. Samples with a probability value greater than 0.5 are classified as positive classes, and samples with a probability value less than 0.5 are classified as negative classes, using 0.5 as a threshold for classification.
And the decision rule is used for judging whether the relay has fault risks according to the analysis result of the feature vector. The implementation process of the decision rule is as follows:
(1) Let N eigenvectors be assumed, each eigenvector containing M eigenvalues, i.e. x= [ X ] 1 ，x 2 ，...，x N ]∈R M ×N Wherein x is i ＝[x i1 ，x i2 ，...，x iM ]Representing an ith feature vector;
(2) According to the characteristic vectors x in the existing K normal states k ＝[x k1 ，x k2 ，...，x k2 ]Calculating the mean vector mu= [ mu ] of each characteristic value under normal state by using a statistical method (such as average value, standard deviation, covariance matrix and the like) 1 ，μ 2 ，...，μ M ]And a covariance matrix Σ;
(3) Based on the existing K eigenvectors x 'with faults' k ＝[x ′k1 ，x′ k2 ，...，x′ kN ]Calculating the mean vector mu '= [ mu ] of each characteristic value under the fault state by using a statistical method' 1 ，μ′ 2 ，...，μ′ M ]And a covariance matrix Σ';
(4) Calculating probability density function p (x|mu,) of feature vector under normal state,probability density function P (x|μ ',Σ') of feature vector under fault condition, prior probability P of normal condition and fault condition normal And P fault ；
(5) For each new feature vector x, calculating the posterior probability P of the feature vector x in the normal state and the fault state according to a Bayesian formula normal (x) And P fault (x) The method comprises the following steps:
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 application.
It will be clear to those skilled in the art that, for convenience and brevity of description, specific working procedures of the above-described systems, apparatuses and units may refer to corresponding procedures in the foregoing method embodiments, and are not repeated herein.
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 each embodiment of the present application 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 foregoing is merely specific embodiments of the present application, but the scope of the present application is not limited thereto, and any person skilled in the art can easily think about changes or substitutions within the technical scope of the present application, and the changes and substitutions are intended to be covered by the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
Finally: the foregoing description of the preferred embodiments of the invention is not intended to limit the invention to the precise form disclosed, and any such modifications, equivalents, and alternatives falling within the spirit and principles of the invention are intended to be included within the scope of the invention.
Claims (8)
1. An analysis system for a non-fault phase impedance relay comprises a data acquisition module, a signal processing module and a decision module, wherein the data acquisition module is connected with the signal processing module, the signal processing module is connected with the decision module, and is characterized in that,
the data acquisition module acquires input and output signals of the relay, converts the signals into acceptable electric signals, and is connected into the data acquisition card, and sequentially performs signal filtering, denoising and phase resistance calculation on the acquired signals; the phase impedance calculation comprises the following specific steps:
step B1, collecting current and voltage signals: collecting current and voltage signals in a power system through a current transformer and a voltage transformer, and obtaining accurate current and voltage waveform data through signal filtering and denoising;
step B2, phase synchronization: the current and voltage signals are subjected to phase synchronization, so that the sampling moments of the current and voltage signals are completely consistent;
step B3, fourier transformation: performing fast Fourier transform on the current and voltage signals, and converting the signals into a frequency domain, wherein the formula is
Wherein X (f) represents a frequency domain signal, N is the number of sampling points, N is the serial number of the sampling points, X (N) is an original signal, f is the frequency, and T is the sampling time interval;
step B4, calculating phase impedance: according to the frequency domain signals of the current and the voltage, the famous Omm law and the kirchhoff law are adopted for calculation to obtain the phase resistance in the power system, and the formula is thatWherein Z (f) is phase resistance, U (f) is a frequency domain signal of voltage, and I (f) is a frequency domain signal of current;
the signal processing module processes and analyzes the signals acquired by the data acquisition module, and a model is built for diagnosing faults through signal analysis and feature extraction;
the decision module is used for predicting and deciding the risk of the working state and faults of the relay based on the data processed by the signal processing module.
2. The analysis system for a non-fault phase impedance relay of claim 1 wherein the denoising process in the data acquisition module comprises the specific steps of:
step A1, frequency domain filtering: performing fast Fourier transform on the acquired data, converting the signal into a frequency domain, and filtering the frequency domain signal by adopting a low-pass, high-pass or band-pass filter according to the signal-to-noise ratio, wherein the formula is X (f) =H (f) Y (f);
wherein X (f) represents the filtered frequency domain signal, H (f) is the frequency response function of the filter, and (f) is the frequency domain representation of the original signal;
step A2, wavelet transformation: wavelet transformation is carried out on the acquired data, signals are decomposed into sub-bands with different scales, noise signals are removed by utilizing the characteristics of the different sub-bands, and the formula is that
Wherein x (n) is an original signal, h (k) and g (k) are wavelet coefficients, and y (n) is a denoised signal;
step A3, statistics of the formulaThe method comprises the following steps: statistical analysis is carried out on the collected data, and denoising processing is carried out on the data through the statistical parameters of mean value, variance, skewness and kurtosis, wherein the formula is that
Wherein,,for the mean value, n is the data length, X i Is the ith data.
3. An analysis system for a non-fault phase impedance relay according to claim 1, wherein the feature extraction method in the signal processing module is: before the feature extraction, the original signal needs to be divided into a time window with a certain length, and then the statistical features of the signal are extracted in the time domain, the frequency spectrum information of the signal is extracted in the frequency domain, and the change information of the signal is extracted in the wavelet domain in sequence.
4. The analysis system for a non-fault phase impedance relay of claim 1 wherein the specific modeling process in the signal processing module is: selecting a model type from the feature data set selected by the feature selection module; setting parameters of the selected model; training the selected model by using a training data set, testing the trained model by using a testing data set, evaluating the performance of the model and optimizing the model, and storing the optimized model into a model library for subsequent use.
5. The analysis system for a non-fault phase impedance relay according to claim 1, wherein the decision module is implemented by a computer or an embedded processor device, and the specific method of the decision module is as follows:
selecting the characteristics obtained by processing by the signal processing module;
based on the feature selection, establishing a corresponding model based on the existing data;
on the basis of establishing a model, corresponding decision rules are determined, and corresponding decisions are carried out according to the result of model prediction.
6. An analysis system for a non-fault phase impedance relay according to claim 5, wherein the specific steps of modeling are:
step C1, preparing a data set, dividing the data set into a training set and a testing set, wherein the training set is used for training a model, and the testing set is used for evaluating the performance of the model;
step C2, adopting a formulaNormalizing the features, wherein x represents the original value of a feature, x min And x max Respectively represent the minimum and maximum values of the feature, x norm Representing the normalized eigenvalue;
step C3, after feature normalization, model training is carried out;
and step C4, after model training is completed, model prediction is performed.
7. The analysis system for a non-fault phase impedance relay according to claim 5, wherein the decision rule is used for judging whether the relay has a fault risk according to the analysis result of the eigenvector; the implementation process of the decision rule is as follows:
(1) Let N eigenvectors be assumed, each eigenvector containing M eigenvalues, i.e. x= [ X ] 1 ，x 2 ，...，x N ]∈R M×N Wherein x is i ＝[x i1 ，x i2 ，...，x iM ]Representing an ith feature vector;
(2) According to the characteristic vectors x in the existing K normal states k ＝[x k1 ，x k2 ，...，x kN ]Calculating the mean vector of each characteristic value under normal state by using a statistical methodμ＝[μ 1 ，μ 2 ，...，μ M ]Sum covariance matrix Σ;
(3) Based on the existing K eigenvectors x 'with faults' k ＝[x ′k1 ，x′ k2 ，...，x′ kN ]Calculating the mean vector mu '= [ mu ] of each characteristic value under the fault state by using a statistical method' 1 ，μ′ 2 ，...，μ′ M ]And a covariance matrix Σ';
(4) Calculating probability density function P (x|mu,) of feature vector in normal state, probability density function P (x|mu', #) of feature vector in fault state, and prior probability P of normal state and fault state normal And P fault ；
(5) For each new feature vector x, calculating the posterior probability P of the feature vector x in the normal state and the fault state according to a Bayesian formula normal () And P fault () The method comprises the following steps:
8. an analysis system for a non-faulted phase impedance relay according to claim 5, wherein the step of feature selection in the information gain:
step S1, calculating the information entropy of each feature, wherein the information entropy is used for measuring the information content of one feature, and the information entropy is calculated by using the following formula:
wherein X represents a feature, p (X i ) Representing that the characteristic X takes the value X i Probability of (2);
step S2, calculating a conditional entropy corresponding to each feature, wherein the conditional entropy is used for measuring the influence degree of one feature on classification, and the conditional entropy is calculated by using the following formula:
wherein Y represents a class label, X represents a feature, p (X i ) Representing that the characteristic X takes the value X i Is H (y|x=x i ) Representing that the characteristic X takes the value X i The conditional entropy of the classification tag Y;
step S3, calculating the information gain of each feature, wherein the information gain is used for measuring the contribution degree of one feature to classification, and the information gain is calculated by using the following formula:
IG(Y，X)＝H(Y)-(Y|X)；
wherein Y represents a classification label, X represents a feature, H (Y) represents information entropy of the classification label Y, and H (Y|X) represents conditional entropy corresponding to the feature X;
and S4, selecting the characteristic with larger information gain as the final characteristic.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310477568.1A CN116482526A (en) | 2023-04-28 | 2023-04-28 | Analysis system for non-fault phase impedance relay |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310477568.1A CN116482526A (en) | 2023-04-28 | 2023-04-28 | Analysis system for non-fault phase impedance relay |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116482526A true CN116482526A (en) | 2023-07-25 |
Family
ID=87222940
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310477568.1A Pending CN116482526A (en) | 2023-04-28 | 2023-04-28 | Analysis system for non-fault phase impedance relay |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116482526A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117992931A (en) * | 2024-04-03 | 2024-05-07 | 国网冀北电力有限公司 | DC signal feature library and reference signal model building system and method |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104712542A (en) * | 2015-01-12 | 2015-06-17 | 北京博华信智科技股份有限公司 | Reciprocating compressor sensitive characteristic extracting and fault diagnosis method based on internet of things |
| CN106772005A (en) * | 2016-12-12 | 2017-05-31 | 深圳市八戒电工技术有限公司 | A kind of relay contact life tests circuit and method |
| US20200200648A1 (en) * | 2018-02-12 | 2020-06-25 | Dalian University Of Technology | Method for Fault Diagnosis of an Aero-engine Rolling Bearing Based on Random Forest of Power Spectrum Entropy |
| CN111650501A (en) * | 2020-03-04 | 2020-09-11 | 苏州热工研究院有限公司 | Testing device for nondestructive online evaluation of aging state of relay |
| CN112363050A (en) * | 2020-09-28 | 2021-02-12 | 广东电网有限责任公司 | SF6 circuit breaker arc contact state evaluation method based on dynamic contact resistance signal |
| CN112629841A (en) * | 2020-11-23 | 2021-04-09 | 江苏科技大学 | Method for predicting service life of relay |
| CN114664605A (en) * | 2020-12-23 | 2022-06-24 | 浙江正泰电器股份有限公司 | Circuit breaker and impedance detection method of contact mechanism thereof |
| CN115728627A (en) * | 2022-10-09 | 2023-03-03 | 上海新联合电气有限公司 | Electric sound contact fault is judgement system in advance |
-
2023
- 2023-04-28 CN CN202310477568.1A patent/CN116482526A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104712542A (en) * | 2015-01-12 | 2015-06-17 | 北京博华信智科技股份有限公司 | Reciprocating compressor sensitive characteristic extracting and fault diagnosis method based on internet of things |
| CN106772005A (en) * | 2016-12-12 | 2017-05-31 | 深圳市八戒电工技术有限公司 | A kind of relay contact life tests circuit and method |
| US20200200648A1 (en) * | 2018-02-12 | 2020-06-25 | Dalian University Of Technology | Method for Fault Diagnosis of an Aero-engine Rolling Bearing Based on Random Forest of Power Spectrum Entropy |
| CN111650501A (en) * | 2020-03-04 | 2020-09-11 | 苏州热工研究院有限公司 | Testing device for nondestructive online evaluation of aging state of relay |
| CN112363050A (en) * | 2020-09-28 | 2021-02-12 | 广东电网有限责任公司 | SF6 circuit breaker arc contact state evaluation method based on dynamic contact resistance signal |
| CN112629841A (en) * | 2020-11-23 | 2021-04-09 | 江苏科技大学 | Method for predicting service life of relay |
| CN114664605A (en) * | 2020-12-23 | 2022-06-24 | 浙江正泰电器股份有限公司 | Circuit breaker and impedance detection method of contact mechanism thereof |
| CN115728627A (en) * | 2022-10-09 | 2023-03-03 | 上海新联合电气有限公司 | Electric sound contact fault is judgement system in advance |
Non-Patent Citations (1)
| Title |
|---|
| 张华歆: "MATLAB图像处理与应用", 安电子科技大学出版社, pages: 267 - 272 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117992931A (en) * | 2024-04-03 | 2024-05-07 | 国网冀北电力有限公司 | DC signal feature library and reference signal model building system and method |
| CN117992931B (en) * | 2024-04-03 | 2024-06-11 | 国网冀北电力有限公司 | DC signal feature library and reference signal model building system and method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111238814B (en) | Rolling bearing fault diagnosis method based on short-time Hilbert transform | |
| CN105841961A (en) | Bearing fault diagnosis method based on Morlet wavelet transformation and convolutional neural network | |
| CN112101174A (en) | LOF-Kurtogram-based mechanical fault diagnosis method | |
| CN111238843B (en) | Fan health evaluation method based on rapid spectrum kurtosis analysis | |
| CN113325277A (en) | Partial discharge processing method | |
| CN116482526A (en) | Analysis system for non-fault phase impedance relay | |
| CN115600088A (en) | Distribution transformer fault diagnosis method based on vibration signals | |
| CN111025100A (en) | Transformer ultrahigh frequency partial discharge signal mode identification method and device | |
| CN108594156B (en) | Improved current transformer saturation characteristic identification method | |
| CN117454081A (en) | Denoising processing method and device for mechanical vibration signal | |
| CN112215307A (en) | Method for automatically detecting signal abnormality of seismic instrument by applying machine learning | |
| CN113495193A (en) | Fault diagnosis method based on multi-wavelet transform and cluster analysis | |
| CN114993671A (en) | Vibration fault diagnosis method and system based on Q factor wavelet transform | |
| CN112213561B (en) | Measurement data preprocessing method and device for leading load parameter noise identification | |
| CN112116917B (en) | Phase jump degree-based method for separating acoustic signals of reactor body and fan | |
| CN112104340B (en) | HMM model and Kalman filtering technology-based switching value input module BIT false alarm reduction method | |
| CN115114955B (en) | Equipment fault detection method based on sound and vibration signals | |
| CN112560674B (en) | Method and system for detecting sound signal quality | |
| CN115659150B (en) | Signal processing method, device and equipment | |
| CN109793511A (en) | Electrocardiosignal noise detection algorithm based on depth learning technology | |
| CN117110989B (en) | Noise fault positioning detection method and system for power equipment | |
| CN117290670B (en) | Transformer bushing insulation state estimation method based on enhancement filter algorithm | |
| CN115545080B (en) | Online detection method and device for lubricating oil scraps | |
| CN112326248B (en) | Based on Hotelling' sT 2 Bearing state monitoring and fault diagnosis method with control chart fused with double kurtosis charts | |
| CN117361256B (en) | Elevator safety management method and system based on artificial intelligence |
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 |