CN114062993A - CVT error state prediction method based on time convolution network - Google Patents

CVT error state prediction method based on time convolution network Download PDF

Info

Publication number
CN114062993A
CN114062993A CN202111218347.XA CN202111218347A CN114062993A CN 114062993 A CN114062993 A CN 114062993A CN 202111218347 A CN202111218347 A CN 202111218347A CN 114062993 A CN114062993 A CN 114062993A
Authority
CN
China
Prior art keywords
time
sequence
data
cvt
convolution network
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
Application number
CN202111218347.XA
Other languages
Chinese (zh)
Inventor
陈文中
陈俊杰
许侃
张金丽
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
State Grid Shanghai Electric Power Co Ltd
East China Power Test and Research Institute Co Ltd
Original Assignee
State Grid Shanghai Electric Power Co Ltd
East China Power Test and Research Institute Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by State Grid Shanghai Electric Power Co Ltd, East China Power Test and Research Institute Co Ltd filed Critical State Grid Shanghai Electric Power Co Ltd
Priority to CN202111218347.XA priority Critical patent/CN114062993A/en
Publication of CN114062993A publication Critical patent/CN114062993A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R35/00—Testing or calibrating of apparatus covered by the other groups of this subclass
    • G01R35/02—Testing or calibrating of apparatus covered by the other groups of this subclass of auxiliary devices, e.g. of instrument transformers according to prescribed transformation ratio, phase angle, or wattage rating
    • 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

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Biomedical Technology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Power Engineering (AREA)
  • Supply And Distribution Of Alternating Current (AREA)

Abstract

The invention relates to a CVT error state prediction method based on a time convolution network. The method comprises the following steps: calculating to obtain CVT characteristic parameter values and generating a time sequence; processing the abnormal value by using a Hampel filter; the processed sequence is checked by using ADF, the stationarity of the processed sequence is judged, and if the time sequence is not stationary, the processed sequence is converted into a stationary time sequence through differential exponential smoothing; training the time convolution network by using the training samples; in the testing stage, inputting a testing sample to the stored network model, outputting an error predicted value of the CVT, comparing the error predicted value with a real value, and checking the effectiveness of the network model; the well-trained time convolution network model is applied to the state prediction of the CVT, and whether the future error of the CVT has the possibility of out-of-tolerance can be judged. The method applies the time convolution network to the state prediction of the CVT, and can obtain the future error state change trend of the CVT so as to check faults in advance and ensure the normal operation of a power grid.

Description

CVT error state prediction method based on time convolution network
Technical Field
The invention relates to a CVT error state prediction method based on a time convolution network, and belongs to the technical field of Capacitor Voltage Transformer (CVT) error state prediction.
Background
The voltage transformer is an important power device in a power grid, converts high voltage into low voltage, and is used for measuring voltage and current values of a primary side. Among them, a Capacitor Voltage Transformer (CVT) has been widely used in 110kV to 500kV power grids, and because of its advantages of good insulating property, simple manufacture, small volume, light weight, etc., the CVT is gradually replacing electromagnetic voltage transformers.
Ideally, the CVT is a time-invariant system, however, over time, the metering characteristics of the CVT can change slowly until the metering error requirements are exceeded. The national standard stipulates that the overhaul time of the CVT does not exceed 4 years. In order to evaluate the operation performance of the CVT, the electric power company usually performs power outage maintenance on the CVT in a certain verification period, and performs error comparison with the CVT to be detected by using a standard. However, the periodic detection has the defects of large workload, low efficiency, untimely detection and the like, and the safe operation of the power system is influenced.
At present, there is an on-line monitoring method for a CVT, which can obtain real-time electrical parameters of the CVT and stably detect the operation performance of the CVT. In practice, the error state of the CVT can be predicted, and the realization of early warning of the system is also important. Error state prediction to the CVT can make the maintenance personal overhaul the CVT before producing the trouble, and the very big degree reduces the electric energy measurement loss, guarantees electric power system's normal operating.
In recent years, artificial intelligence technology is mature, and prediction of data by using a machine learning method is widely applied. The time convolution network can process the problem of time sequence, the reception field can be increased by increasing the convolution kernel, so that a high layer can fuse a large amount of bottom layer information, and the time sequence data processing method has better precision and processing speed for processing long time span. The historical running error state data of the system is utilized and trained through the time convolution network, the future running error state of the system can be predicted, and whether the possibility of error out-of-tolerance exists in the future or not is judged.
Disclosure of Invention
The purpose of the invention is: the system can predict the future system operation error state according to the historical operation error state data, and judge whether the CVT has the possibility of error out-of-tolerance.
In order to achieve the above object, the technical solution of the present invention is to provide a CVT error state prediction method based on a time convolution network, which is characterized by comprising the following steps:
s1, collecting electric parameters of the capacitor voltage transformer during operation, and calculating to obtain four characteristic parameters of the capacitor voltage transformer, namely amplitude error epsilonUPhase error phiUDielectric loss tan delta, capacitance value C, amplitude error epsilonUAnd phase error phiUThe accuracy of the capacitor voltage transformer is reflected, and the dielectric loss tan delta and the capacitance value C reflect the safety and reliability of the capacitor voltage transformer; taking the measured values of the characteristic parameters at certain time intervals as sequence values to generate a time sequence of four characteristic parameters of the capacitor voltage transformer;
s2, eliminating abnormal values deviating from most characteristic values and/or not conforming to the statistical rule of the measured data in each time sequence by using a Hampel filtering method, so as to avoid the influence of abnormal values possibly generated when the parameters of the capacitive voltage transformer are measured on the prediction precision, which results in error of the error state prediction result;
s3, checking the filtered time sequence by using ADF to judge the stationarity, if the time sequence is not stationary, carrying out differential exponential smoothing on the time sequence to convert the time sequence into a stationary time sequence, and if all the time sequences are stationary, entering the step S4;
s4, normalizing the time sequence obtained in the last step;
s5, training the time convolution network by using the time sequence after normalization processing obtained from the steps S1 to S4 based on the sample data to obtain the time convolution network parameter with the most accurate prediction result;
and S5, applying the well-trained time convolution network to error prediction of the capacitor voltage transformer, and inputting the time sequence obtained by normalization processing through the steps S1 to S4 based on real-time data into the time convolution network to obtain the future error state change trend of the capacitor voltage transformer so as to check faults in advance and ensure normal operation of a power grid.
Preferably, in the step S1, the amplitude error eUIs the ratio of the primary voltage and the secondary output voltage of the capacitor voltage transformer and is expressed as
Figure BDA0003311559110000021
Wherein, KrFor rated transformation ratio, U1Is the primary voltage value, U, of a capacitive voltage transformer2Measuring the voltage for the second time of the capacitor voltage transformer;
phase error phiUThe difference value of the secondary voltage phase and the primary voltage phase of the capacitor voltage transformer is represented as phiU＝Φ2-Φ1Wherein phi is1Is the phase of the primary voltage, [ phi ]2Is the phase of the secondary voltage when phi2Ahead of phi1Time, phase error phiUIs a positive value;
the dielectric loss tan delta is expressed as
Figure BDA0003311559110000022
Wherein phiICThe phase of the leakage current of the capacitor voltage transformer is shown;
a capacitance value C is expressed as
Figure BDA0003311559110000023
ICThe leakage current of the capacitor voltage transformer is shown, and omega is the frequency of the power grid.
Preferably, in the step S2, the time sequence x isnThe detection and deletion of the abnormal value by using the Hampel filter specifically comprises the following steps:
s21, for time series xnArbitrary ith data xiCalculating data xiAnd time series xnMedian m of six data around itiSuppose data xiAnd time series xnWherein the six surrounding data form a sequence Xi＝{xi-3，xi-2，xi-1，xi，xi+1，xi+2，xi+3Is then mi＝median(Xi) Wherein, mean (X)i) Is a sequence XiA median of (d);
s22, estimating data x by using median absolute deviationiStandard deviation of median Si，Si＝d×median(|xi-3-mi|，…|xi+3-mi|), wherein d is an unbiased estimate of the gaussian distribution, and the specific expression is:
Figure BDA0003311559110000031
where erfc (x) is a complementary error function;
s23, if the data xiIf the difference from the median is less than 3 times the standard deviation, the data xiNot an abnormal value, data xiRemain unchanged, otherwise, if data xiIf the difference from the median is greater than 3 times the standard deviation, the data xiFor abnormal values, use the median miReplacement data xi。
Preferably, the step S3 specifically includes the following steps:
s31, using ADF to check the stability of the time sequence, and completing the steps by the following three models:
model one:
Figure BDA0003311559110000032
model two:
Figure BDA0003311559110000033
and (3) model III:
Figure BDA0003311559110000034
in the formula, xtRepresenting a time series xnMiddle t data, Δ xt＝xt-xt-1Denotes the difference between adjacent data in time series, lambdaiDenotes the autoregressive coefficient, utRepresenting a mean of 0 and a variance of σ2When d is equal to 0, the sequence has a unit root and is a non-stationary sequence, and when d is less than 0, the sequence has no unit root and is a stationary sequence;
if the test of one model in the first model, the second model and the third model rejects the null hypothesis, the current time sequence has no unit root and is a stable sequence, and the step S4 is entered, otherwise, the test is continued until the current time sequence traverses all the models, if the original hypothesis is not rejected, the current time sequence is a non-stable sequence, and the step S32 is entered;
s32, smoothing the current time sequence by using a differential exponential smoothing method:
if the current time sequence is in a growing trend, a stable new sequence is formed by using a first-order difference exponential smoothing model:
Figure BDA0003311559110000041
in the formula (I), the compound is shown in the specification,
Figure BDA0003311559110000042
a predicted value at time t, i.e., a smoothed value,
Figure BDA0003311559110000043
Represents a first order difference exponential smoothing value, alpha represents a smoothing constant, and the value range is [0, 1 ]]；
If the current time sequence shows a quadratic growth trend, a stable new sequence is formed by utilizing a second-order difference exponential smoothing model:
Figure BDA0003311559110000044
Figure BDA0003311559110000045
in the formula (I), the compound is shown in the specification,
Figure BDA0003311559110000046
representing a second order difference;
if the current time sequence shows an n-time increasing trend, a stable new sequence is formed by using an n-order differential exponential smoothing model, wherein n is more than or equal to 2.
Preferably, in the step S4, the normalizing the time series specifically includes the following steps:
using Min-Max method to process ith data x in time sequenceiNormalization processing is carried out to obtain normalized data
Figure BDA0003311559110000047
Figure BDA0003311559110000048
The calculation formula is as follows:
Figure BDA0003311559110000049
xminand xmaxRespectively the minimum value and the maximum value in the current time series.
Preferably, the step S5 specifically includes the following steps,
s51, dividing the data into training samples and testing samples, wherein the first 80% are training samples, and the second 20% are testing samples;
s52, training by using the training sample as the input of the time convolution network;
s53, judging whether the training error meets the expectation, if so, saving the network and entering the step S54, otherwise, further adjusting the parameters of the time convolution network such as weight, threshold value, convolution value and the like;
and S54, inputting the normalized test sample by using the stored time convolution network structure to obtain a prediction result, comparing the prediction result with a real prediction value, judging whether the test accuracy meets the preset requirement, if so, storing the network model for predicting the error state of the capacitor voltage transformer, otherwise, returning to the step S53, and resetting parameters such as the size of a convolution kernel, the expansion coefficient and the like.
The invention provides a CVT error state prediction method based on a time convolution network, which is characterized in that after a time sequence is subjected to abnormal value elimination and smoothing processing, the time convolution network is used for training, so that a system can predict a future system operation error state according to historical operation error state data, and whether the CVT has the possibility of error overproof is judged.
Due to the adoption of the technical scheme, compared with the prior art, the invention has the following advantages and positive effects: the invention eliminates abnormal values of the time sequence of the characteristic parameters, and avoids the influence of the abnormal values on the prediction precision due to sensor noise, channel interference and the like to the maximum extent. And the sequence is subjected to smooth processing, so that the prediction result is not influenced by non-smooth behaviors such as seasonal variation, cyclic variation and the like. In practice, the error state prediction of the CVT usually needs to predict the state of 24 hours or days in the future, and the method uses the time convolution network to predict the error state of the CVT in the future, so that the prediction accuracy and the speed are greatly improved when long-time span time series data are processed.
Drawings
FIG. 1 is a flow chart of a CVT error state prediction method of the present invention;
fig. 2 is a flow chart of the CVT characteristic parameter time series pre-processing normalization of the present invention.
Detailed Description
The invention is further illustrated by the following specific examples. It should be understood that these examples are for illustrative purposes only and are not intended to limit the scope of the present invention. Further, it should be understood that various changes or modifications of the present invention may be made by those skilled in the art after reading the teaching of the present invention, and such equivalents may fall within the scope of the present invention as defined in the appended claims.
The embodiment discloses a CVT error state prediction method based on a time convolution network, which comprises the following steps as shown in figure 1:
1) collecting electric parameters of the CVT during operation, and calculating to obtain four characteristic parameters of the CVT, namely amplitude error epsilonUPhase error phiUDielectric loss tan δ and capacitance C. For the characteristic parameters of the CVT, taking the measured values at certain time intervals as sequence values to generate a time sequence of the characteristic parameters of the CVT;
2) using Hampel filtering to remove most of the characteristic values deviating from the time sequence of each characteristic parameter value and abnormal values which do not accord with the statistical rule of the measured data;
3) using ADF to check the processed sequence, judging the stationarity of the processed sequence, and if the time sequence is not stationary, performing differential exponential smoothing on the processed sequence to convert the time sequence into a stationary time sequence;
4) normalizing the time sequence, training the time convolution network according to the sample data, and obtaining the time convolution network parameter with the most accurate prediction result;
5) the well-trained time convolution network is applied to error prediction of the CVT, and the future error state change trend of the CVT can be obtained, so that faults can be checked in advance, and normal operation of a power grid is guaranteed.
Compared with other CVT error state prediction methods, the method has the advantages that the future error state of the CVT is predicted by using the time convolution network, the future error state of the CVT is predicted according to the time sequence of the historical characteristic parameter values of the CVT, and the method has good accuracy and speed in processing long time span time sequence data.
It is intended that the foregoing detailed description be regarded as illustrative rather than limiting, and that it be understood that it is the following claims, including all equivalents, that are intended to define the spirit and scope of this invention. The above examples are to be construed as merely illustrative and not limitative of the remainder of the disclosure. After reading the description of the invention, the skilled person can make various changes or modifications to the invention, and these equivalent changes and modifications also fall into the scope of the invention defined by the claims.

Claims (6)

1. A CVT error state prediction method based on a time convolution network is characterized by comprising the following steps:
s1, collecting electric parameters of the capacitor voltage transformer during operation, and calculating to obtain four characteristic parameters of the capacitor voltage transformer, namely amplitude error epsilonUPhase error phiUDielectric loss tan delta, capacitance value C, amplitude error epsilonUAnd phase error phiUThe accuracy of the capacitor voltage transformer is reflected, and the dielectric loss tan delta and the capacitance value C reflect the safety and reliability of the capacitor voltage transformer; taking the measured values of the characteristic parameters at certain time intervals as sequence values to generate a time sequence of four characteristic parameters of the capacitor voltage transformer;
s2, eliminating abnormal values deviating from most characteristic values and/or not conforming to the statistical rule of the measured data in each time sequence by using a Hampel filtering method, so as to avoid the influence of abnormal values possibly generated when the parameters of the capacitive voltage transformer are measured on the prediction precision, which results in error of the error state prediction result;
s3, checking the filtered time sequence by using ADF to judge the stationarity, if the time sequence is not stationary, carrying out differential exponential smoothing on the time sequence to convert the time sequence into a stationary time sequence, and if all the time sequences are stationary, entering the step S4;
s4, normalizing the time sequence obtained in the last step;
s5, training the time convolution network by using the time sequence after normalization processing obtained from the steps S1 to S4 based on the sample data to obtain the time convolution network parameter with the most accurate prediction result;
and S5, applying the well-trained time convolution network to error prediction of the capacitor voltage transformer, and inputting the time sequence obtained by normalization processing through the steps S1 to S4 based on real-time data into the time convolution network to obtain the future error state change trend of the capacitor voltage transformer so as to check faults in advance and ensure normal operation of a power grid.
2. The CVT error state prediction method based on time convolution network of claim 1 wherein in step S1, the magnitude error εUIs the ratio of the primary voltage and the secondary output voltage of the capacitor voltage transformer and is expressed as
Figure FDA0003311559100000011
Wherein, KrFor rated transformation ratio, U1Is the primary voltage value, U, of a capacitive voltage transformer2Measuring the voltage for the second time of the capacitor voltage transformer;
phase error phiUThe difference value of the secondary voltage phase and the primary voltage phase of the capacitor voltage transformer is represented as phiU＝Φ2-Φ1Wherein phi is1Is the phase of the primary voltage, [ phi ]2Is the phase of the secondary voltage when phi2Ahead of phi1Time, phase error phiUIs a positive value;
the dielectric loss tan delta is expressed as
Figure FDA0003311559100000012
Wherein phiICThe phase of the leakage current of the capacitor voltage transformer is shown;
a capacitance value C is expressed as
Figure FDA0003311559100000021
ICThe leakage current of the capacitor voltage transformer is shown, and omega is the frequency of the power grid.
3. The CVT error state prediction method based on time convolution network of claim 1 wherein in step S2, for time series xnThe detection and deletion of the abnormal value by using the Hampel filter specifically comprises the following steps:
s21, for time series xnArbitrary ith data xiCalculating data xiAnd time series xnMedian m of six data around itiSuppose data xiAnd time series xnWherein the six surrounding data form a sequence Xi＝{xi-3，xi-2，xi-1，xi，xi+1，xi+2，xi+3Is then mi＝median(Xi) Wherein, mean (X)i) Is a sequence XiA median of (d);
s22, estimating data x by using median absolute deviationiStandard deviation of median Si，Si＝d×median(|xi-3-mi|，…|xi+3-mi|), wherein d is an unbiased estimate of the gaussian distribution, and the specific expression is:
Figure FDA0003311559100000022
where erfc (x) is a complementary error function;
s23, if the data xiIf the difference from the median is less than 3 times the standard deviation, the data xiNot an abnormal value, data xiRemain unchanged, otherwise, if data xiIf the difference from the median is greater than 3 times the standard deviation, the data xiFor abnormal values, use the median miReplacement data xi。
4. The CVT error state prediction method based on the time-convolution network as recited in claim 1, wherein the step S3 specifically includes the following steps:
s31, using ADF to check the stability of the time sequence, and completing the steps by the following three models:
model one:
Figure FDA0003311559100000023
model two:
Figure FDA0003311559100000024
and (3) model III:
Figure FDA0003311559100000025
in the formula, xtRepresenting a time series xnMiddle t data, Δ xt＝xt-xt-1Denotes the difference between adjacent data in time series, lambdaiDenotes the autoregressive coefficient, utRepresenting a mean of 0 and a variance of σ2When d is equal to 0, the sequence has a unit root and is a non-stationary sequence, and when d is less than 0, the sequence has no unit root and is a stationary sequence;
if the test of one model in the first model, the second model and the third model rejects the null hypothesis, the current time sequence has no unit root and is a stable sequence, and the step S4 is entered, otherwise, the test is continued until the current time sequence traverses all the models, if the original hypothesis is not rejected, the current time sequence is a non-stable sequence, and the step S32 is entered;
s32, smoothing the current time sequence by using a differential exponential smoothing method:
if the current time sequence is in a growing trend, a stable new sequence is formed by using a first-order difference exponential smoothing model:
Figure FDA0003311559100000031
in the formula (I), the compound is shown in the specification,
Figure FDA0003311559100000032
a predicted value at time t, i.e., a smoothed value,
Figure FDA0003311559100000033
Represents a first order difference exponential smoothing value, alpha represents a smoothing constant, and the value range is [0, 1 ]]；
If the current time sequence shows a quadratic growth trend, a second-order difference exponential smoothing model is used for forming a smoothnessThe new sequence of (2):
Figure FDA0003311559100000034
Figure FDA0003311559100000035
in the formula (I), the compound is shown in the specification,
Figure FDA0003311559100000036
representing a second order difference;
if the current time sequence shows an n-time increasing trend, a stable new sequence is formed by using an n-order differential exponential smoothing model, wherein n is more than or equal to 2.
5. The CVT error state prediction method based on the time-convolution network as claimed in claim 1, wherein the step S4 of normalizing the time series specifically includes the steps of:
using Min-Max method to process ith data x in time sequenceiNormalization processing is carried out to obtain normalized data
Figure FDA0003311559100000037
Figure FDA0003311559100000038
The calculation formula is as follows:
Figure FDA0003311559100000039
xminand xmaxRespectively the minimum value and the maximum value in the current time series.
6. The CVT error state prediction method based on time convolution network of claim 1 wherein the step S5 includes the following steps,
s51, dividing the data into training samples and testing samples, wherein the first 80% are training samples, and the second 20% are testing samples;
s52, training by using the training sample as the input of the time convolution network;
s53, judging whether the training error meets the expectation, if yes, saving the network and entering the step S54, otherwise, further adjusting the parameters of the time convolution network;
and S54, inputting the normalized test sample by using the stored time convolution network structure to obtain a prediction result, comparing the prediction result with a real prediction value, judging whether the test accuracy meets the preset requirement, if so, storing the network model for predicting the error state of the capacitor voltage transformer, otherwise, returning to the step S53, and resetting the parameters.
CN202111218347.XA 2021-10-20 2021-10-20 CVT error state prediction method based on time convolution network Pending CN114062993A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111218347.XA CN114062993A (en) 2021-10-20 2021-10-20 CVT error state prediction method based on time convolution network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111218347.XA CN114062993A (en) 2021-10-20 2021-10-20 CVT error state prediction method based on time convolution network

Publications (1)

Publication Number Publication Date
CN114062993A true CN114062993A (en) 2022-02-18

Family

ID=80235232

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111218347.XA Pending CN114062993A (en) 2021-10-20 2021-10-20 CVT error state prediction method based on time convolution network

Country Status (1)

Country Link
CN (1) CN114062993A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114710218A (en) * 2022-05-31 2022-07-05 深圳市佳贤通信设备有限公司 Distributed node and base station communication efficiency optimization method based on 5G
CN114818817A (en) * 2022-05-06 2022-07-29 国网四川省电力公司电力科学研究院 Weak fault recognition system and method for capacitive voltage transformer
CN114938339A (en) * 2022-05-19 2022-08-23 中国农业银行股份有限公司 Data processing method and related device

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114818817A (en) * 2022-05-06 2022-07-29 国网四川省电力公司电力科学研究院 Weak fault recognition system and method for capacitive voltage transformer
CN114818817B (en) * 2022-05-06 2023-05-19 国网四川省电力公司电力科学研究院 Weak fault identification system and method for capacitive voltage transformer
CN114938339A (en) * 2022-05-19 2022-08-23 中国农业银行股份有限公司 Data processing method and related device
CN114710218A (en) * 2022-05-31 2022-07-05 深圳市佳贤通信设备有限公司 Distributed node and base station communication efficiency optimization method based on 5G

Similar Documents

Publication Publication Date Title
CN116702081B (en) Intelligent inspection method for power distribution equipment based on artificial intelligence
CN114062993A (en) CVT error state prediction method based on time convolution network
CN111474510A (en) Error evaluation method and system for voltage transformer with non-stable output
CN110763997A (en) Early fault early warning method for synchronous motor stator
CN106529124A (en) A transformer insulation state assessment method based on principal component analysis and support vector machines
CN112069727B (en) Intelligent transient stability evaluation system and method with high reliability for power system
CN114460445B (en) Transformer aging unavailability evaluation method considering aging threshold and service life
CN113325357A (en) Voltage transformer error evaluation method and system based on output time series difference
CN114740418A (en) Error online detection method for capacitor voltage transformer
CN116960941A (en) Power climbing event prediction method, device, equipment and storage medium
CN114372344B (en) Quantitative identification method for subsynchronous oscillation damping characteristic influence factors
CN117349614A (en) Frequency stability prediction method based on self-attention mechanism and space-time diagram convolution network
CN116662771A (en) Method and system for self-adaptive PCA error quantitative evaluation of capacitive voltage transformer
Pandit et al. Using Gaussian process theory for wind turbine power curve analysis with emphasis on the confidence intervals
CN103928923B (en) A kind of network stationary power quality method for early warning based on sensitivity analysis
CN109635430A (en) Grid power transmission route transient signal monitoring method and system
CN112098066A (en) High-voltage shunt reactor fault diagnosis method and system based on gate control circulation unit
CN116226620A (en) Subsynchronous oscillation mode identification method based on SVM and LSTM
CN116401545A (en) Multimode fusion type turbine runout analysis method
Wilson et al. Uncertainty Quantification of Capacitor Switching Transient Location Using Machine Learning
CN115423370A (en) Relay protection equipment health state assessment method and device
CN105656453B (en) A kind of optical fiber current mutual inductor random noise Real-Time Filtering method based on time series
CN109787219B (en) Intelligent identification method for voltage sag
LIU et al. An application of multistep long-short term memory algorithm to predict the performance of capacitor voltage transformer
Zhang et al. Research on Data-driven Self-diagnosis for Measurement Errors in Capacitor Voltage Transformers

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