CN113609765B - Overvoltage prediction method - Google Patents
Overvoltage prediction method Download PDFInfo
- Publication number
- CN113609765B CN113609765B CN202110865043.6A CN202110865043A CN113609765B CN 113609765 B CN113609765 B CN 113609765B CN 202110865043 A CN202110865043 A CN 202110865043A CN 113609765 B CN113609765 B CN 113609765B
- Authority
- CN
- China
- Prior art keywords
- overvoltage
- factors
- influence
- factor
- model
- 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
- 238000000034 method Methods 0.000 title claims abstract description 26
- 238000004088 simulation Methods 0.000 claims abstract description 16
- 238000012545 processing Methods 0.000 claims abstract description 10
- 230000008569 process Effects 0.000 claims abstract description 9
- 238000012847 principal component analysis method Methods 0.000 claims abstract description 3
- 239000011159 matrix material Substances 0.000 claims description 16
- 238000012549 training Methods 0.000 claims description 14
- 239000003990 capacitor Substances 0.000 claims description 12
- 230000001186 cumulative effect Effects 0.000 claims description 8
- 238000004458 analytical method Methods 0.000 claims description 7
- 238000009413 insulation Methods 0.000 claims description 6
- 238000013507 mapping Methods 0.000 claims description 6
- 210000002569 neuron Anatomy 0.000 claims description 6
- 238000013528 artificial neural network Methods 0.000 claims description 5
- 238000009826 distribution Methods 0.000 claims description 3
- 238000011478 gradient descent method Methods 0.000 claims description 3
- 230000016507 interphase Effects 0.000 claims description 3
- 238000012544 monitoring process Methods 0.000 claims description 3
- 230000007935 neutral effect Effects 0.000 claims description 3
- 230000010355 oscillation Effects 0.000 claims description 3
- 230000000694 effects Effects 0.000 claims description 2
- 238000000342 Monte Carlo simulation Methods 0.000 claims 1
- 238000012360 testing method Methods 0.000 claims 1
- 230000000630 rising effect Effects 0.000 abstract 1
- 230000006870 function Effects 0.000 description 10
- 238000010586 diagram Methods 0.000 description 5
- 238000005259 measurement Methods 0.000 description 4
- 238000011084 recovery Methods 0.000 description 3
- 238000011156 evaluation Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000010219 correlation analysis Methods 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 230000001939 inductive effect Effects 0.000 description 1
- 230000005389 magnetism Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
- 238000009827 uniform distribution Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- 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
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/10—Numerical modelling
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Biology (AREA)
- Medical Informatics (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention discloses an overvoltage prediction method, which comprises the following steps: s1, building an overvoltage fault tree model and analyzing overvoltage influence factors; s2, constructing an influence factor randomness overvoltage model and analyzing influence; s3, establishing an overvoltage prediction model considering random factors and processing influence factors; s4, determining and identifying key factors; and S5, solving the prediction model to finally obtain a prediction result with higher accuracy. According to the invention, various factors influencing the breaking overvoltage are analyzed by using a fault tree, an ATP simulation model is built for breaking process simulation, on the basis of obtaining overvoltage data, the influence factors are independently checked by using a principal component analysis method, and the stray capacitance, line parameters, the insulating strength rising rate of a vacuum circuit breaker and the breaking moment are 4 main influence factors.
Description
Technical Field
The invention relates to the technical field of overvoltage measurement, in particular to an overvoltage prediction method.
Background
Overvoltage refers to the phenomenon of long-time voltage variation that the root mean square value of alternating current voltage is increased by more than 10% of rated value under power frequency and the duration is longer than 1 minute, and the overvoltage is usually the instant result of load switching and occurs under the condition of switching on or off of inductive or capacitive load in normal use.
Various interference factors occur during overvoltage measurement, and different factors have great influence on the measurement value of the overvoltage.
Therefore, it is necessary to invent an overvoltage prediction method to solve the above problems.
Disclosure of Invention
In view of the shortcomings in the above-mentioned problems, the present invention provides an overvoltage prediction method.
In order to achieve the above object, the present invention provides the following solutions:
an overvoltage prediction method, comprising the steps of:
s1, building an overvoltage fault tree model and analyzing overvoltage influence factors;
s2, constructing an influence factor randomness overvoltage model and analyzing influence;
s3, establishing an overvoltage prediction model considering random factors and processing influence factors;
s4, determining and identifying key factors;
and S5, solving the prediction model to finally obtain a prediction result with higher accuracy.
Further, the step S1 is to build an overvoltage fault tree model and analyze influence factors in the overvoltage influence factors, and the step is to:
s101, building an overvoltage fault tree model;
s102, determining influence factors through a voltage fault tree model, wherein the influence factors are classified into two main types:
a. the TRV is an oscillation result of charge and discharge between the reactor and the capacitor in the switching-on and switching-off process of the circuit breaker from the energy perspective, the capacitor comprises an inter-phase capacitor, a relative ground capacitor and the like, and because two ends of the circuit breaker contact are respectively connected with a power supply side and a reactor side, the TRV is affected by the parameters of the system power supply and the reactor together, including parameters such as a power supply phase angle, a voltage amplitude, a line parameter, a reactor capacity, a reactor remanence and the like in switching;
b. the factor of the insulation voltage.
Further, the step of constructing an influence factor stochastic overvoltage model and the influence analysis step in the step S2 are as follows:
s201, building an influence factor randomness overvoltage model;
s202, determining the value range of each influence factor aiming at a certain system, carrying out characteristic value on each influence factor, and carrying out normalization processing on the influence factors with different dimensions to enable the influence factors to be located in a [0,1] interval;
s203, setting an overvoltage interval threshold value, intercepting an overvoltage section, determining the value of each influence factor in overvoltage, analyzing an overvoltage waveform, extracting overvoltage characteristics and constructing a mapping set with the corresponding influence factor mapping relation;
s204, analyzing probability distribution of each factor, simulating the value of each influence factor Monte Carlo, building a simulation model of the system, inversely normalizing the Monte Carlo simulated value, and sending the value into the simulation model to obtain a system overvoltage waveform;
s205, analyzing and extracting the amplitude and duration of the overvoltage and the corresponding influence factor value.
Further, the step of establishing the overvoltage prediction model considering the random factor in S3 includes the following steps:
s301, constructing an overvoltage model considering randomness of influence factors, and defining an F-representing influence factor matrix as assuming that m factors influencing the amplitude and duration of overvoltage exist
O(V,T)=[F1,F2,......Fm]
Wherein V, T respectively represent the overvoltage times and the corresponding durations; fi (i=1, 2..n.) shows column vector of the i-th influencing factor, length is N, analyze the relation between the influencing factors to ensure mutual independence between the factors, for this purpose, use the result of clustering analysis to make independence check, if the influencing factors are independent of each other, make prediction, otherwise, need decorrelation, after the influencing factors are independent of each other, use the principal component method to determine the principal factor among many influencing factors, namely
Representing n samples, m influencing factors, and recording F' as normalized factor matrix, wherein the principal component Z after PCA processing is
Wherein, Z= { Z1, Z2,..zp } is p (p.ltoreq.m) principal components, yij is the coefficient of the jth factor of the ith principal component;
s302, carrying out data standardization on F to obtain F 'so that the average value of F' becomes zero;
s303, solving a covariance matrix of F, solving a eigenvalue of the covariance matrix, and arranging eigenvectors according to the eigenvalue lambdaj to form a matrix P;
s304, obtaining the dimension-reduced data Y by calculating Y=PF';
s305, calculating a contribution rate alpha i and a cumulative contribution rate beta i of each feature root by the following formula:
and selecting the main component according to the principle that the cumulative contribution rate beta i is more than 85 percent.
Further, the step of determining and identifying the S4 key factor is as follows:
s401, building an over-voltage prediction model based on an ANN, wherein the over-voltage prediction model of the ANN is divided into an input layer, an hidden layer and an output layer, each layer is formed by interconnecting a large number of nodes, each node represents a specific output function, data is input from the input layer, a result is transmitted through the node as the next layer of input through the operation of the node, the node simulates a biological signal transmitting function in a neuron, a predicted value is finally output, the predicted value is compared with a set threshold, the deviation between the predicted value and the set threshold is defined as a loss function, the size and the positive and negative of the loss function are the basis for adjusting the nodes of the simulated neuron, the power saving function of the neural network is adjusted, and the predicted result is globally searched by combining a gradient descent method or a conjugate descent method so as to achieve the predicted result with the minimum predicted deviation.
Further, in the step S5, a prediction model is solved, a result is obtained according to the substation neutral point current monitoring data, error comparison is continuously performed on the result and an actual value, if the error cannot meet the requirement, the retraining is performed again until the error reaches the standard, the accuracy of the training result is achieved, the bias current levels under different combinations of influence factors can be estimated through the obtained prediction model, the bias current levels are used as training data again, the prediction model is formed through training, and finally the prediction result with higher accuracy is obtained.
The beneficial effects of the invention are as follows: according to the invention, various factors influencing the break-over voltage are analyzed by using a fault tree, an ATP simulation model is built for carrying out break-over process simulation, on the basis of obtaining overvoltage data, the influence factors are independently checked by using a principal analysis method, 4 principal influence factors including stray capacitance, line parameters, the rise rate of the insulation strength of a vacuum circuit breaker and the break-over time are found, the factors are used as the input of a neural network for training and predicting, and compared with a simulation result, the error of the obtained predicted data is not more than 15%, and the overvoltage condition can be predicted more reliably.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a diagram of an on-off overvoltage simulation model of a reactor according to the present invention;
fig. 2 is a waveform diagram of TRV when the circuit breaker of the present invention is opened;
FIG. 3 is a graph showing the arc current waveform when the circuit breaker of the present invention is opened;
FIG. 4 is a graph of the contribution rate and cumulative contribution rate of the various factors of the present invention;
FIG. 5 is a graph showing comparison between ANN prediction and ATP simulation results according to the present invention;
FIG. 6 is a graph of ANN versus ATP error according to the present invention;
FIG. 7 is a trend chart of the training convergence process of the present invention;
FIG. 8 is a graph of the trend of data fitting evaluation index change in the training process of the invention;
FIG. 9 is a diagram of a shunt reactor switching overvoltage fault tree of the present invention;
FIG. 10 is a schematic diagram of a shunt reactor switching overvoltage equivalent circuit according to the present invention;
fig. 11 is a schematic flow chart of a circuit breaker opening process according to the present invention;
fig. 12 is a diagram of a prediction model of the switching overvoltage ANN of the vacuum circuit breaker according to the present invention.
Detailed Description
The invention will be further described with reference to the drawings and detailed description. Example 1
An overvoltage prediction method as shown in fig. 1 to 12, comprising the steps of:
s1, building an overvoltage fault tree model and analyzing overvoltage influence factors;
s2, constructing an influence factor randomness overvoltage model and analyzing influence;
s3, establishing an overvoltage prediction model considering random factors and processing influence factors;
s4, determining and identifying key factors;
and S5, solving the prediction model to finally obtain a prediction result with higher accuracy.
As shown in fig. 9, the step S1 of building an overvoltage fault tree model and analyzing the influence factors in the overvoltage influence factors includes:
s101, building an overvoltage fault tree model;
s102, determining influence factors through a voltage fault tree model, wherein the influence factors are classified into two main types:
a. the TRV factor is from the energy perspective, the TRV factor is an oscillation result of charge and discharge between the reactor and the capacitor in the switching-on and switching-off process of the circuit breaker, the capacitor comprises an inter-phase capacitor and a relative ground capacitor, and because two ends of the circuit breaker contact are respectively connected with a power supply side and a reactor side, the TRV is affected by the common effects of the system power supply and the reactor parameters, including a power supply phase angle, a voltage amplitude, a line parameter, a reactor capacity and a reactor remanence parameter during switching;
b. the factor of the insulation voltage.
As shown in fig. 10-11, the step of constructing the influence factor stochastic overvoltage model and the influence analysis in S2 is as follows:
s201, building an influence factor randomness overvoltage model;
s202, determining the value range of each influence factor aiming at a certain system, carrying out characteristic value on each influence factor, and carrying out normalization processing on the influence factors with different dimensions to enable the influence factors to be located in a [0,1] interval;
s203, setting an overvoltage interval threshold value, intercepting an overvoltage section, determining the value of each influence factor in overvoltage, analyzing an overvoltage waveform, extracting overvoltage characteristics and constructing a mapping set with the corresponding influence factor mapping relation;
s204, analyzing probability distribution of each factor, simulating the value of each influence factor Monte Carlo, building a simulation model of the system, inversely normalizing the Monte Carlo simulated value, and sending the value into the simulation model to obtain a system overvoltage waveform;
s205, analyzing and extracting the amplitude value and duration of the overvoltage and the corresponding influence factor value thereof; the step of establishing the overvoltage prediction model considering the random factors and the influence factors in the step S3 is as follows:
s301, constructing an overvoltage model considering randomness of influence factors, and defining an F-representing influence factor matrix as assuming that m factors influencing the amplitude and duration of overvoltage exist
Wherein V, T respectively represent the overvoltage times and the corresponding durations; fi (i=1, 2..n.) shows column vector of the i-th influencing factor, length is N, analyze the relation between the influencing factors to ensure mutual independence between the factors, for this purpose, use the result of clustering analysis to make independence check, if the influencing factors are independent of each other, make prediction, otherwise, need decorrelation, after the influencing factors are independent of each other, use the principal component method to determine the principal factor among many influencing factors, namely
Representing n samples, m influencing factors, and recording F' as normalized factor matrix, wherein the principal component Z after PCA processing is
Wherein, Z= { Z1, Z2,..zp } is p (p.ltoreq.m) principal components, yij is the coefficient of the jth factor of the ith principal component;
s302, carrying out data standardization on F to obtain F 'so that the average value of F' becomes zero;
s303, solving a covariance matrix of F, solving a eigenvalue of the covariance matrix, and arranging eigenvectors according to the eigenvalue lambdaj to form a matrix P;
s304, obtaining the dimension-reduced data Y by calculating Y=PF';
s305, calculating a contribution rate alpha i and a cumulative contribution rate beta i of each feature root by the following formula:
and selecting the main component according to the principle that the cumulative contribution rate beta i is more than 85 percent.
As shown in fig. 12, the steps of determining and identifying the S4 key factors are as follows: s401, building an over-voltage prediction model based on an ANN, wherein the over-voltage prediction model of the ANN is divided into an input layer, an hidden layer and an output layer, a large number of nodes in each layer are connected with each other, each node represents a specific output function, data is input from the input layer, a result is transmitted through the node as the next layer through the operation of the node, the node simulates a biological signal transmitting function in a neuron, a predicted value is finally output, the predicted value is compared with a set threshold, the deviation between the predicted value and the set threshold is defined as a loss function, the size and the positive and negative of the loss function are the basis for adjusting the nodes of the simulated neuron, the power saving function of the neural network is adjusted, and the predicted result is globally searched by combining a gradient descent method or a conjugate descent method and the like so as to achieve the predicted result with the minimum predicted deviation; and S5, solving a predictive model, obtaining a result according to the current monitoring data of the neutral point of the transformer substation, carrying out error comparison on the result and an actual value continuously, if the error cannot meet the requirement, returning to retraining until the error reaches the standard, estimating bias current levels under different combinations of influence factors through the obtained predictive model, taking the bias current levels as training data again, forming a predictive model continuously through training, and finally obtaining a predictive result with higher accuracy.
Example 1:
the embodiment of the invention provides a method for identifying an influence factor of overvoltage measurement, which comprises the following steps:
in the 500kV ultra-high voltage system shown in fig. 1, each parameter represents symbol which is marked in the figure, wherein a breaker control module can adjust the breaking time of a breaker, and simulate the action characteristic of the vacuum breaker according to the response of the system, the initial phase angle of a power supply in a simulation model is set to be minus 30 degrees, the inductance of the power supply of the system is 548mH, the length of a transmission line is 300km, a JMRT model is selected, three phases are simultaneously broken, a transient recovery voltage waveform (TRV) of the breaker is selected, the cut-off voltage is 5A, the part of a curve surrounded by a current waveform broken line is amplified when the breaker is broken, and the current flowing in the circuit breaker indicates that an arc reburning phenomenon occurs due to breakdown of an insulating medium between contacts caused by charge and discharge overvoltage of a reactor.
Factors influencing the breaking reactor are classified into stray capacitance (denoted as x 1), line parameters (denoted as x 2) and electricity
The residual magnetism of the resistor (marked as x 3), the line inductance (marked as x 4), the switching-off time (marked as x 5), the insulation recovery voltage of the circuit breaker (marked as x 6), the connection line parameter (marked as x 7) and the capacity L2 of the reactor (marked as x 8), wherein the switching-on time of the circuit breaker is controlled to be normally distributed near 0.01s, the deviation is 0.0004, 1000 discrete points are uniformly and randomly distributed at the switching-on moment, the initial phase angle of the power supply takes values in a range of 0-180 degrees, and 1.3p.u is taken as an overvoltage threshold value. Selecting a value of +/-50% from the vicinity of the standard value by adopting a uniform distribution mode for the rest parameters;
to facilitate correlation analysis, the influence factors were normalized and then examined for independence, with results shown in Table 1
As shown in table 1, the influence factors are not completely independent, so that decorrelation processing by a principal component analysis method is also required;
determination of the correlation coefficient matrix of Table 1 eigenvalue λi (i=) 1, 2....8.). Obtaining characteristic values of all influence factors as {2.31,2.13,0.22,0.35,2.66,2.19,0.82,0.18}, and calculating contribution rates alpha and accumulated contribution rates beta of all the factors;
as can be seen from fig. 4, the cumulative contribution rate of the principal components of the 1 st, 2 nd, 5 th and 6 th influencing factors reaches 86% >85%, so that these 4 factors can be selected as inputs of the prediction model;
the training period is set to 35000, the prediction precision is set to 1e-3, the initial learning rate is 0.1, self-adaptive adjustment can be carried out in the training process, and the training sample predicts the overvoltage and the error thereof;
5-6, comparing the ANN neural network prediction and ATP simulation results with errors under the same working condition;
7-8, it can be seen that using ANN prediction can ensure that the overvoltage predicted value deviation is controlled within 15%, wherein the deviation of the overvoltage peak value is not more than 10%;
as can be seen from fig. 7, the prediction model can be close to the target value in about 35000 cycles of training, the mean square error is less than 0.001, the corresponding learning rate is 0.302, the data fitting evaluation index shows that the fitting degree between the prediction value and the simulation data is higher, the regression coefficient of the prediction model reaches 0.9994, the prediction result is more reliable, and the stray capacitance, the line parameter, the opening time and the insulation recovery voltage of the circuit breaker are also indicated as key influence factors of overvoltage.
Claims (3)
1. An overvoltage prediction method, comprising the steps of:
s1, building an overvoltage fault tree model and analyzing overvoltage influence factors;
s2, constructing an influence factor randomness overvoltage model and analyzing influence;
and (2) constructing an influence factor randomness overvoltage model and an influence analysis step as follows:
s201, building an influence factor randomness overvoltage model;
s202, aiming at a certain system, determining the value range of each influence factor, carrying out characteristic value on each influence factor, and carrying out characteristic value on each influence factor
Normalizing the influence factors of the same dimension to enable the influence factors to be located in a [0,1] interval;
s203, setting an overvoltage interval threshold value, intercepting an overvoltage section, determining the value of each influence factor in overvoltage, and
analyzing the overvoltage waveform, extracting overvoltage characteristics and mapping relation with corresponding influence factors, and constructing a mapping set;
s204, analyzing probability distribution of each factor, and simulating and taking value of each influence factor Monte Carlo, and building simulation of the system
The true model is used for reversely normalizing the numerical value of Monte Carlo simulation and then sending the numerical value into the simulation model to obtain a system overvoltage waveform;
s205, analyzing and extracting the amplitude value and duration of the overvoltage and the corresponding influence factor value thereof;
s3, establishing an overvoltage prediction model considering random factors and processing influence factors;
the step of establishing the overvoltage prediction model considering the random factors and the influence factors in the step S3 is as follows:
s301, constructing an overvoltage model considering influence factor randomness, and supposing that m influence overvoltage amplitudes and durations exist
Defining F as the factor of the influence factor matrix
O(V,T)=[F1,F2,......Fm]
Wherein V, T respectively represent the overvoltage multiple and the corresponding duration;
fi (i=1, 2..n.) shows the column vector of the i-th influencing factor, the length is N, the relation between the influencing factors is analyzed to ensure the mutual independence of the factors, the result of clustering analysis can be used for the independence test, if the influencing factors are mutually independent, the prediction can be performed, otherwise, the decorrelation is needed, after the factors are mutually independent, the principal component analysis method can be used for determining the principal factors in a plurality of influencing factors, namely
Representing n samples, m influencing factors, and recording F' as normalized factor matrix, wherein the principal component Z after PCA processing is
Wherein, Z= { Z1, Z2,..zp } is p (p.ltoreq.m) principal components, yij is the coefficient of the jth factor of the ith principal component;
s302, carrying out data standardization on F to obtain F 'so that the average value of F' becomes zero;
s303, solving a covariance matrix of F, solving a eigenvalue of the covariance matrix, and arranging eigenvectors according to the eigenvalue lambdaj to form a matrix P;
s304, obtaining the dimension-reduced data Y by calculating Y=PF';
s305, calculating a contribution rate alpha i and a cumulative contribution rate beta i of each feature root by the following formula:
according to the cumulative contribution rate beta>The principle of 85 percent is that the main component is selected;
s4, determining and identifying key factors;
the step of determining and identifying the S4 key factors is as follows:
s401, building an over-voltage prediction model based on an ANN, wherein the over-voltage prediction model of the ANN is divided into an input layer, an hidden layer and an output layer, a large number of nodes in each layer are interconnected to form the over-voltage prediction model, each node represents a specific output function, data is input from the input layer, the result is transmitted through the nodes as the next layer input through the operation of the nodes, and the nodes simulate biological signals
In the neuron propagation function, finally outputting a predicted value, comparing the predicted value with a set threshold value, defining the deviation between the predicted value and the set threshold value as a loss function, adjusting the power saving function of the neural network by taking the magnitude and the positive and negative of the loss function as the basis for adjusting the simulated neuron nodes, and carrying out global search on the predicted result by combining a gradient descent method or a conjugate descent method so as to achieve the predicted result with the minimum predicted deviation;
and S5, solving the prediction model to finally obtain a prediction result with higher accuracy.
2. The method for predicting overvoltage according to claim 1, wherein the step S1 of constructing an overvoltage fault tree model and analyzing the influence factors of the overvoltage influence factors comprises the steps of:
s101, building an overvoltage fault tree model;
s102, determining influence factors through a voltage fault tree model, wherein the influence factors are classified into two main types:
a. the TRV factor is from the energy perspective, the TRV factor is an oscillation result of charge and discharge between the reactor and the capacitor in the switching-on and switching-off process of the circuit breaker, the capacitor comprises an inter-phase capacitor and a relative ground capacitor, and because two ends of the circuit breaker contact are respectively connected with a power supply side and a reactor side, the TRV is affected by the common effects of the system power supply and the reactor parameters, including a power supply phase angle, a voltage amplitude, a line parameter, a reactor capacity and a reactor remanence parameter during switching;
b. the factor of the insulation voltage.
3. The method for predicting overvoltage according to claim 1, wherein the step S5 is to solve a prediction model, obtain a result according to the substation neutral point current monitoring data, compare the result with the actual value continuously, and if the error fails to meet the requirement, return to retraining until the error reaches the standard, estimate bias current levels under different combinations of influencing factors by using the obtained prediction model, and use the bias current levels as training data again, form a prediction model by training, and finally obtain a prediction result with higher accuracy.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110865043.6A CN113609765B (en) | 2021-07-29 | 2021-07-29 | Overvoltage prediction method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110865043.6A CN113609765B (en) | 2021-07-29 | 2021-07-29 | Overvoltage prediction method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113609765A CN113609765A (en) | 2021-11-05 |
CN113609765B true CN113609765B (en) | 2023-12-26 |
Family
ID=78306018
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110865043.6A Active CN113609765B (en) | 2021-07-29 | 2021-07-29 | Overvoltage prediction method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113609765B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117093929B (en) * | 2023-07-06 | 2024-03-29 | 珠海市伊特高科技有限公司 | Cut-off overvoltage prediction method and device based on unsupervised domain self-adaptive network |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104217104A (en) * | 2014-08-19 | 2014-12-17 | 上海交通大学 | Power transformer service life analysis method and system based on risk evaluation |
EP3101570A1 (en) * | 2015-06-04 | 2016-12-07 | The MathWorks, Inc. | Extension of model-based design to identify and analyze impact of reliability information on systems and components |
CN109271975A (en) * | 2018-11-19 | 2019-01-25 | 燕山大学 | A kind of electrical energy power quality disturbance recognition methods based on big data multi-feature extraction synergetic classification |
CN109617046A (en) * | 2018-11-22 | 2019-04-12 | 中国电力科学研究院有限公司 | Power distribution network method for analyzing stability in a kind of electric system |
CN110233476A (en) * | 2019-04-09 | 2019-09-13 | 广东电网有限责任公司 | Voltage stability assessment method and relevant apparatus during a kind of black starting-up |
CN110689195A (en) * | 2019-09-26 | 2020-01-14 | 云南电网有限责任公司电力科学研究院 | Power daily load prediction method |
CN111429034A (en) * | 2020-04-21 | 2020-07-17 | 国网信通亿力科技有限责任公司 | Method for predicting power distribution network fault |
CN112000923A (en) * | 2020-07-14 | 2020-11-27 | 中国电力科学研究院有限公司 | Power grid fault diagnosis method, system and equipment |
-
2021
- 2021-07-29 CN CN202110865043.6A patent/CN113609765B/en active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104217104A (en) * | 2014-08-19 | 2014-12-17 | 上海交通大学 | Power transformer service life analysis method and system based on risk evaluation |
EP3101570A1 (en) * | 2015-06-04 | 2016-12-07 | The MathWorks, Inc. | Extension of model-based design to identify and analyze impact of reliability information on systems and components |
CN109271975A (en) * | 2018-11-19 | 2019-01-25 | 燕山大学 | A kind of electrical energy power quality disturbance recognition methods based on big data multi-feature extraction synergetic classification |
CN109617046A (en) * | 2018-11-22 | 2019-04-12 | 中国电力科学研究院有限公司 | Power distribution network method for analyzing stability in a kind of electric system |
CN110233476A (en) * | 2019-04-09 | 2019-09-13 | 广东电网有限责任公司 | Voltage stability assessment method and relevant apparatus during a kind of black starting-up |
CN110689195A (en) * | 2019-09-26 | 2020-01-14 | 云南电网有限责任公司电力科学研究院 | Power daily load prediction method |
CN111429034A (en) * | 2020-04-21 | 2020-07-17 | 国网信通亿力科技有限责任公司 | Method for predicting power distribution network fault |
CN112000923A (en) * | 2020-07-14 | 2020-11-27 | 中国电力科学研究院有限公司 | Power grid fault diagnosis method, system and equipment |
Non-Patent Citations (5)
Title |
---|
基于故障树法和层次分析法的电力变压器状态综合评估;姚峰;张忠会;张毅明;孙建华;谢义苗;何乐彰;;电气应用(11);P70-73 * |
基于模糊层次分析法的变压器状态评估;张晶晶;许修乐;丁明;李金忠;王健一;吴超;;电力系统保护与控制(03);P80-86 * |
基于蒙特卡罗仿真的电力变压器故障树分析;康新兴;;国外电子测量技术(10);P5-9 * |
小电流接地系统铁磁谐振过电压关键影响因素辨识;何龙;马金财;杜龙基;李军;吴伟丽;刘勇;;电气工程学报(02);P61-69 * |
自适应神经模糊网络高压输电线路操作过电压风险点定位算法;杨虎臣;王晓东;安慧;李毅靖;郑钟;林永春;;中国测试(07);P27-35 * |
Also Published As
Publication number | Publication date |
---|---|
CN113609765A (en) | 2021-11-05 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Cavraro et al. | Power distribution network topology detection with time-series signature verification method | |
CN106980905B (en) | Power distribution network power supply reliability prediction method and system | |
Lotfifard et al. | A systematic approach for ranking distribution systems fault location algorithms and eliminating false estimates | |
CN111814390B (en) | Voltage transformer error prediction method based on transfer entropy and wavelet neural network | |
Lira et al. | MOSA monitoring technique based on analysis of total leakage current | |
Jiang et al. | Stochastic subspace identification‐based approach for tracking inter‐area oscillatory modes in bulk power system utilising synchrophasor measurements | |
CN106485089B (en) | The interval parameter acquisition methods of harmonic wave user's typical condition | |
CN113447783B (en) | Voltage transformer insulation fault identification model construction method and device | |
CN112149873B (en) | Low-voltage station line loss reasonable interval prediction method based on deep learning | |
CN109270482B (en) | Online evaluation method for metering accuracy of voltage transformer and terminal equipment | |
CN111025222A (en) | Electronic voltage transformer state monitoring method based on recursion principal component analysis | |
CN107831415A (en) | The Interval Valued Fuzzy diversity method that a kind of transformer insulating paper ageing state is assessed | |
CN113609765B (en) | Overvoltage prediction method | |
CN117310353B (en) | Method and system for testing through-flow pressurization faults of primary and secondary circuits of transformer substation | |
CN106408016A (en) | Distribution network power outage time automatic identification model construction method | |
CN109033498A (en) | Winding deformation recognition methods based on transmission function feature principal component and neural network | |
CN108416148A (en) | A kind of high-altitude electromagnetic pulse field wire coupling uncertainty acquisition methods based on polynomial chaos expression | |
CN116840764A (en) | Method and system for evaluating comprehensive error state of capacitive voltage transformer | |
Shukla et al. | Enhancing the reliability of six‐phase transmission line protection using power quality informatics with real‐time validation | |
Niazazari et al. | Event cause analysis in distribution networks using synchro waveform measurements | |
CN115586402B (en) | Power distribution network fault diagnosis and processing method | |
CN117630543A (en) | Transformer fault state supervision method and system based on Internet of things data | |
CN112710923A (en) | Data-driven single-phase earth fault line selection method based on post-fault steady-state information | |
Chen et al. | A neural network approach for evaluating distribution system reliability | |
Shi et al. | Travelling waves‐based identification of sub‐health condition of feeders in power distribution system |
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 |