CN115392141A - Self-adaptive current transformer error evaluation method - Google Patents

Self-adaptive current transformer error evaluation method Download PDF

Info

Publication number
CN115392141A
CN115392141A CN202211331032.0A CN202211331032A CN115392141A CN 115392141 A CN115392141 A CN 115392141A CN 202211331032 A CN202211331032 A CN 202211331032A CN 115392141 A CN115392141 A CN 115392141A
Authority
CN
China
Prior art keywords
error
value
current
evaluation
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.)
Granted
Application number
CN202211331032.0A
Other languages
Chinese (zh)
Other versions
CN115392141B (en
Inventor
饶芳
陈勉舟
陈应林
代洁
任波
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.)
Wuhan Gelanruo Intelligent Technology Co ltd
Original Assignee
Wuhan Gelanruo Intelligent Technology 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 Wuhan Gelanruo Intelligent Technology Co ltd filed Critical Wuhan Gelanruo Intelligent Technology Co ltd
Priority to CN202211331032.0A priority Critical patent/CN115392141B/en
Publication of CN115392141A publication Critical patent/CN115392141A/en
Application granted granted Critical
Publication of CN115392141B publication Critical patent/CN115392141B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

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
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods

Abstract

The invention relates to a self-adaptive current transformer error evaluation method, which comprises the following steps: taking each line CT on the same bus in the same transformer substation as an evaluation group, constructing a physical relation between each CT individual error and a measured value node current vector sum, obtaining a measurement error of each CT in the group, and recording the measurement error as a first error estimation value
Figure 93888DEST_PATH_IMAGE001
(ii) a Establishing an individual CT error evaluation model; obtaining a second error estimation value of each CT to be estimated in the estimation group based on the individual CT error estimation model; fusing the first error estimation value and the second error estimation value by adopting an improved physical information neural network method to obtain a third error estimation value; estimating the first error and the second error by using the error deviation valueAnd the difference estimation value is corrected, so that real-time online evaluation of the CT operation error is realized in a self-adaptive manner.

Description

Self-adaptive current transformer error evaluation method
Technical Field
The invention relates to the field of smart power grids, in particular to a self-adaptive current transformer error evaluation method.
Background
Current transformers (Current transformers) are important measurement devices in electrical power systems. The primary winding is connected in series in the main loop of the power transmission and transformation, and the secondary winding is respectively connected with equipment such as a measuring instrument, a relay protection or an automatic device and the like according to different requirements, and is used for changing the large current of the primary loop into the small current of the secondary side for the measurement and control protection metering equipment to safely collect. The method is accurate and reliable, and has great significance for safe operation, control protection, electric energy metering and trade settlement of the power system.
At present, the current transformer error evaluation usually adopts an off-line checking method or an on-line checking method, and the ratio difference and the angle difference of the current transformer are obtained through a direct comparison method. However, these methods have long verification period, complicated field wiring and low working efficiency. In order to perfect a current transformer error state evaluation system, a current transformer error state evaluation method needs to be established urgently, the problem that the error of the current transformer is out of tolerance is found, the out-of-limit operation time of the current transformer error is reduced, and the transformer detection work is guided, so that the fairness of electric energy metering is ensured.
Disclosure of Invention
The invention provides a self-adaptive current transformer error evaluation method aiming at the technical problems in the prior art, and the real-time online evaluation of CT operation errors is self-adaptively realized according to different CT electrical physical relations of a transformer substation by utilizing a multi-dimensional feature extraction and integrated fusion learning method.
According to a first aspect of the present invention, there is provided an adaptive current transformer error evaluation method, comprising: step 1, taking each line CT on the same bus in the same transformer substation as an evaluation group S, constructing a physical relation between each CT individual error and a current vector sum of a measured value node according to a kirchhoff current law, obtaining a measurement error of each CT in the group by using a mathematical processing method, and recording the measurement errorFirst error estimate for CT
Figure 879023DEST_PATH_IMAGE001
；
Step 2, establishing an individual CT error evaluation model by using an off-line modeling and on-line training mode; the input of the individual CT error evaluation model is a multi-dimensional characteristic parameter of CT, and the output is a second error estimation value of CT
Figure 669125DEST_PATH_IMAGE002
(ii) a Obtaining a second error estimation value of each CT to be evaluated in the evaluation group S based on the individual CT error evaluation model
Figure 456077DEST_PATH_IMAGE002
；
Step 3, adopting an improved physical information neural network method to estimate the first error value
Figure 724248DEST_PATH_IMAGE001
And said second error estimate
Figure 778791DEST_PATH_IMAGE002
Performing fusion to obtain a third error estimation value
Figure 739794DEST_PATH_IMAGE003
；
Step 4, respectively calculating a first error estimation value
Figure 246999DEST_PATH_IMAGE001
The second error estimation value
Figure 335172DEST_PATH_IMAGE002
And a third error estimate
Figure 542424DEST_PATH_IMAGE003
Error deviation value therebetween
Figure 205487DEST_PATH_IMAGE004
And
Figure 465567DEST_PATH_IMAGE005
(ii) a Using said error deviation value
Figure 75540DEST_PATH_IMAGE004
And
Figure 104676DEST_PATH_IMAGE005
for the first error estimation value
Figure 141902DEST_PATH_IMAGE001
And said second error estimate
Figure 889278DEST_PATH_IMAGE002
And correcting and adaptively realizing the online evaluation of the CT running error.
On the basis of the technical scheme, the invention can be improved as follows.
Optionally, the step 1 includes:
step 101, acquiring and screening current measurement values of stable sections of CT of each line under the same node of a transformer substation in real time, and constructing a monitoring data set;
102, constructing a relational expression between a current measurement value and an error of a target CT in the monitoring data set and current measurement values and errors of other CTs according to a kirchhoff current law;
step 103, calculating a current true value of the target CT by using the target phase current data and the rated transformation ratio of the target CT as a current reference value
Figure 798547DEST_PATH_IMAGE006
(ii) a Selecting three-phase current data of the CT of any line in the monitoring data set for state evaluation, and respectively taking the corresponding CT and phase when the current data is in a normal state as a target CT and a target phase;
step 104, taking the current measurement values of other CTs except the target CT in the monitoring data set as input, and taking the current reference value
Figure 416610DEST_PATH_IMAGE006
As output, training the LAPO-RBF neural network to obtain a LAPO-RBF neural network model;
105, obtaining the current reference value of each to-be-evaluated CT by using the LAPO-RBF neural network model, and calculating a first error estimation value of each to-be-evaluated CT according to the current reference value
Figure 155896DEST_PATH_IMAGE001
。
Optionally, in the step 102
The obtained relation is:
Figure 859410DEST_PATH_IMAGE007
(10)
Figure 607923DEST_PATH_IMAGE008
(11)
Figure 611651DEST_PATH_IMAGE009
、
Figure 446140DEST_PATH_IMAGE010
approximately 0,n is the number of lines, m represents A, B or the C-phase,
Figure 902529DEST_PATH_IMAGE011
is the real part of the current of the measured value,
Figure 153601DEST_PATH_IMAGE012
is the imaginary part of the current of the measured value,
Figure 11835DEST_PATH_IMAGE013
for each of the line current measurements, the line current measurement,
Figure 827345DEST_PATH_IMAGE014
for the purpose of the corresponding ratio difference,
Figure 771030DEST_PATH_IMAGE015
corresponding phase differences.
Optionally, the multidimensional feature parameter in step 2 includes: working condition parameters, electromagnetic parameters and environmental parameters.
Optionally, step 2 includes:
step 201, acquiring multi-dimensional characteristic parameters of each CT in an off-line detection process to construct an off-line data set; training an ANN model based on the offline data set to obtain a pre-training model;
step 202, acquiring multi-dimensional characteristic parameters of each CT in the online detection process, and constructing each segmented data set according to the current value range of the CT;
step 203, inputting each segmented data set into each pre-training model for training, and obtaining each individual CT error evaluation model corresponding to the current value range of each CT;
step 204, inputting the online monitoring data of the CT to be evaluated into the individual CT error evaluation model corresponding to the segment where the current value is positioned, and outputting a second error estimation value of the CT to be evaluated by the individual CT error evaluation model
Figure 64608DEST_PATH_IMAGE002
。
Optionally, in the step 3, a fusion model of the error evaluation value is obtained by using an improved physical information neural network method;
the input characteristic quantity of the fusion model is as follows:
Figure 308507DEST_PATH_IMAGE016
the output characteristic quantity is:
Figure 294918DEST_PATH_IMAGE017
；
Figure 86419DEST_PATH_IMAGE003
is the third error estimate for the CT to be evaluated.
Optionally, the construction process of the fusion model includes:
approximating a function through a fully-connected neural network, solving partial differential equation residual error and initial value residual error constraints by using an automatic differential technology, putting the partial differential equation residual error and initial value residual error constraints into a loss function as regular terms, and obtaining a neural network connection weight parameter W and a partial differential equation physical parameter b by using a chameleon algorithm.
Optionally, the first error estimation value in the step 4 is used
Figure 386950DEST_PATH_IMAGE001
And said second error estimate
Figure 845875DEST_PATH_IMAGE002
The process of making the correction includes:
when the group evaluation condition is met, obtaining the error evaluation value of the CT to be evaluated
Figure 534346DEST_PATH_IMAGE018
；
When the population evaluation condition is not met, the error evaluation value of the CT to be evaluated is obtained
Figure 452623DEST_PATH_IMAGE019
。
According to the self-adaptive current transformer error evaluation method provided by the invention, the thought of multi-dimensional feature extraction and integrated fusion learning is utilized, and the physical information neural network algorithm after optimization of the chameleon algorithm is adopted, so that the complementation and mutual correction of the evaluation results of small-scale groups (single-group CT) and large-scale groups (3 groups and above) are realized, and the accuracy and the adaptive range of CT error evaluation are improved.
Drawings
FIG. 1 is a flow chart of a method for error estimation of an adaptive current transformer according to the present invention;
FIG. 2 (a) is a schematic diagram of a single-phase current transformer;
FIG. 2 (b) is an equivalent circuit diagram of a single-phase current transformer;
fig. 3 is a network structure diagram of a physical information neural network.
Detailed Description
The principles and features of this invention are described below in conjunction with the following drawings, which are set forth by way of illustration only and are not intended to limit the scope of the invention.
Fig. 1 is a flowchart of an adaptive current transformer error evaluation method provided by the present invention, and as shown in fig. 1, the method includes:
step 1, taking each line CT (current transformer) on the same bus in the same transformer substation as an evaluation group S, constructing a physical relation between each CT individual error and a current vector sum of a measured value node according to a kirchhoff current law, obtaining a measurement error of each CT in the group by using a mathematical processing method, and recording the measurement error as a first error estimation value of the CT
Figure 822425DEST_PATH_IMAGE001
。
Step 2, establishing an individual CT error evaluation model by using an off-line modeling and on-line training mode; the input of the individual CT error evaluation model is the multi-dimensional characteristic parameters of CT, and the output is the second error estimation value of CT
Figure 306496DEST_PATH_IMAGE002
(ii) a Obtaining a second error estimation value of each CT to be estimated in the estimation group S based on the individual CT error estimation model
Figure 634709DEST_PATH_IMAGE002
。
Step 3, adopting an improved physical information neural network (Floyd-PINNs) method to estimate the first error value
Figure 276168DEST_PATH_IMAGE020
And a second error estimate
Figure 980819DEST_PATH_IMAGE002
Performing fusion to obtain a third error estimation value
Figure 522659DEST_PATH_IMAGE003
。
Step (ii) of4, respectively calculating the first error estimated values
Figure 756194DEST_PATH_IMAGE001
The second error estimation value
Figure 914643DEST_PATH_IMAGE002
And a third error estimate
Figure 688564DEST_PATH_IMAGE003
Error deviation value therebetween
Figure 111674DEST_PATH_IMAGE004
And
Figure 578427DEST_PATH_IMAGE005
(ii) a Using error deviation values
Figure 896276DEST_PATH_IMAGE004
And
Figure 473888DEST_PATH_IMAGE005
for the first error estimation value
Figure 724741DEST_PATH_IMAGE001
And a second error estimate
Figure 598281DEST_PATH_IMAGE002
And correcting and adaptively realizing the online evaluation of the CT running error.
The invention provides a self-adaptive current transformer error evaluation method, which utilizes a method of multi-dimensional feature extraction and integrated fusion learning to realize real-time online evaluation of CT operation errors in a self-adaptive manner according to different CT electrical and physical relationships of a transformer substation.
Example 1
Embodiment 1 provided by the present invention is an embodiment of an adaptive current transformer error evaluation method provided by the present invention, and as can be seen in conjunction with fig. 1, the embodiment of the error evaluation method includes:
step 1, providing electricity in the same transformer substationThe CT of the physical relationship is constructed into a group S', each line CT on the same bus is used as an evaluation group S, the physical relationship between individual errors of each CT and the sum of current vectors of measurement value nodes is constructed according to the kirchhoff current law, the measurement error of each CT in the group is obtained by using a mathematical processing method (a least square method, a neural network method and the like), and the measurement error is recorded as a first error estimation value of the CT
Figure 200164DEST_PATH_IMAGE001
。
In one possible embodiment, step 1 includes:
step 101, obtaining and screening current measurement values of stable sections of CT of each line under the same node of the transformer substation in real time, and constructing a monitoring data set.
And step 102, constructing a relational expression between the current measurement value and the error of the target CT in the monitoring data set and the current measurement values and the errors of other CTs according to the kirchhoff current law.
In one possible embodiment, step 102
As shown in fig. 2 (a) and fig. 2 (b), which are schematic diagrams of a single-phase current transformer and equivalent circuit diagrams thereof, respectively, it can be known from fig. 2 (a) and fig. 2 (b):
the rated current ratio is:
Figure 784729DEST_PATH_IMAGE021
(1)
using magnetic potential to balance:
Figure 155667DEST_PATH_IMAGE022
(2)
due to the existence of
Figure 167486DEST_PATH_IMAGE023
The difference between the primary current value and the secondary current value is divided into a ratio error
Figure 991085DEST_PATH_IMAGE024
Phase difference with
Figure 379341DEST_PATH_IMAGE025
As shown in formulas (3) and (4):
Figure 604786DEST_PATH_IMAGE026
(3)
Figure 288971DEST_PATH_IMAGE027
(4)
in the formula (I), the compound is shown in the specification,
Figure 865445DEST_PATH_IMAGE028
is the rated transformation ratio of the CT,
Figure 791813DEST_PATH_IMAGE029
and
Figure 871765DEST_PATH_IMAGE030
respectively the amplitude and phase of the actual primary current of the CT,
Figure 490965DEST_PATH_IMAGE031
and
Figure 289156DEST_PATH_IMAGE032
respectively are applied under the measuring conditions
Figure 753636DEST_PATH_IMAGE033
The amplitude and phase of the actual secondary current.
Selecting all incoming and outgoing lines on the same bus as a group, wherein in the group:
Figure 953673DEST_PATH_IMAGE034
for each true value of the line current,
Figure 245239DEST_PATH_IMAGE035
the magnitude of the true value of the current is,
Figure 530727DEST_PATH_IMAGE036
the phase of the true value of the current is,
Figure 64477DEST_PATH_IMAGE037
for each of the line current measurements,
Figure 119020DEST_PATH_IMAGE038
for the purpose of the corresponding ratio difference,
Figure 814444DEST_PATH_IMAGE039
for the corresponding phase difference, n lines are provided.
Figure 587228DEST_PATH_IMAGE040
(5)
Figure 659089DEST_PATH_IMAGE041
(6)
Figure 568139DEST_PATH_IMAGE042
(7)
Figure 31586DEST_PATH_IMAGE043
(8)
Figure 26087DEST_PATH_IMAGE044
(9)
The resulting relationship is:
Figure 901639DEST_PATH_IMAGE045
(10)
Figure 930775DEST_PATH_IMAGE008
(11)
Figure 968001DEST_PATH_IMAGE046
、
Figure 715377DEST_PATH_IMAGE047
approximately 0,n is the number of lines, m represents A, B or the C phase,
Figure 129041DEST_PATH_IMAGE048
is the real part of the current of the measured value,
Figure 12684DEST_PATH_IMAGE049
is the imaginary part of the current of the measured value,
Figure 424073DEST_PATH_IMAGE050
for each of the line current measurements,
Figure 894631DEST_PATH_IMAGE051
for the purpose of the corresponding ratio difference,
Figure 111986DEST_PATH_IMAGE015
the corresponding phase difference.
There is the formula available: and the current and the error value of each line are constructed into a functional relation, wherein the measured value and the current of each line are known quantities, and the error is unknown quantity.
Step 103, calculating the current true value of the target CT by using the target phase current data and the rated transformation ratio of the target CT as the current reference value
Figure 850135DEST_PATH_IMAGE006
(ii) a And selecting three-phase current data of the CT of any line in the monitoring data set for state evaluation, and respectively taking the corresponding CT and phase as a target CT and a target phase when the current data is in a normal state.
Step 104, taking the current measurement values of other CT except the target CT in the monitoring data set as input, and taking the current reference value
Figure 494743DEST_PATH_IMAGE006
As output, the LAPO-RBF neural network is trained to obtain LAPAnd (4) an O-RBF neural network model.
105, obtaining a current reference value of each CT to be evaluated by using the LAPO-RBF neural network model, and calculating a first error estimation value of each CT to be evaluated according to the current reference value
Figure 951132DEST_PATH_IMAGE001
。
Step 2, establishing an individual CT error evaluation model by using an off-line modeling and on-line training mode; the input of the individual CT error evaluation model is the multi-dimensional characteristic parameter of CT, and the output is the second error estimation value of CT
Figure 706598DEST_PATH_IMAGE002
(ii) a Obtaining a second error estimation value of each CT to be estimated in the estimation group S based on the individual CT error estimation model
Figure 361571DEST_PATH_IMAGE002
。
In a possible embodiment, the multidimensional characteristic parameter in step 2 includes: working condition parameters, electromagnetic parameters and environmental parameters.
In one possible embodiment, step 2 includes:
step 201, acquiring multi-dimensional characteristic parameters of each CT in an off-line detection process to construct an off-line data set; and training an ANN model based on the offline data set to obtain a pre-training model.
Step 202, acquiring multi-dimensional characteristic parameters of each CT in the online detection process, and constructing each segmented data set according to the current value range of the CT through data preprocessing.
That is, each current value range is set, and the offline data set is divided into each segmented data set corresponding to each current value range.
Step 203, inputting each segmented data set into each pre-training model respectively for training to obtain each individual CT error evaluation model corresponding to the range of the current value of each CT;
step 204, inputting the online monitoring data of the CT to be evaluated into the individual C corresponding to the section where the current value is positionedA T error evaluation model, wherein the individual CT error evaluation model outputs a second error estimation value of the CT to be evaluated
Figure 380342DEST_PATH_IMAGE002
。
Step 3, adopting an improved physical information neural network (Floyd-PINNs) method to estimate the first error value
Figure 356651DEST_PATH_IMAGE001
And a second error estimate
Figure 915808DEST_PATH_IMAGE002
Performing fusion to obtain a third error estimation value
Figure 628549DEST_PATH_IMAGE003
。
In a possible embodiment mode, a fusion model of the error estimation value is obtained in step 3 by adopting a method of improving a physical information neural network.
The input characteristic quantity of the fusion model is as follows:
Figure 614960DEST_PATH_IMAGE016
the output characteristic quantity is:
Figure 45941DEST_PATH_IMAGE017
；
Figure 143210DEST_PATH_IMAGE003
is the third error estimate for the CT to be evaluated.
In a possible embodiment, the construction process of the fusion model includes:
approximating a function by a fully-connected neural network, solving partial differential equation residual errors and initial value residual error constraints by using an automatic differential technology, putting the partial differential equation residual errors and the initial value residual error constraints into a loss function as a regular term, and obtaining a neural network connection weight parameter W and partial differential equation physical parameters b by using a chameleon algorithm.
Specifically, as shown in fig. 3, a network structure diagram of the physical information neural network (pins) is shown, and with reference to fig. 3, the principle of the physical information neural network (pins) is as follows: after time and space data are input, a function is approximated through a fully-connected neural network, then constraint of partial differential equation residual and initial value residual is solved by using an automatic differential technology and is put into a loss function as a regular term, and finally a neural network connection weight parameter (W) and a partial differential equation physical parameter (b) are obtained by using a gradient descent method.
Because the gradient descent can not ensure the convergence to the global optimal solution, the invention optimizes the hyper-parameters W and b of the physical information neural network by using the chameleon algorithm (Floyd algorithm) and constructs the improved physical information neural network (Floyd-PINNs).
The chameleon algorithm comprises the following steps:
1) And initializing hyper-parameters W and b of the physical information neural network.
2) And searching for prey.
3) The color-changing longan is rotated.
4) And capturing prey.
The first error estimation value
Figure 710458DEST_PATH_IMAGE001
And a second error estimate
Figure 369234DEST_PATH_IMAGE002
Solving third error estimation value of CT to be evaluated by substituting error estimation value fusion model
Figure 21932DEST_PATH_IMAGE003
。
Step 4, respectively calculating first error estimated values
Figure 922892DEST_PATH_IMAGE001
The second error estimation value
Figure 610226DEST_PATH_IMAGE002
And a third error estimate
Figure 938439DEST_PATH_IMAGE003
Error deviation value therebetween
Figure 78433DEST_PATH_IMAGE004
And
Figure 251925DEST_PATH_IMAGE005
(ii) a Using error deviation values
Figure 793765DEST_PATH_IMAGE004
And
Figure 788485DEST_PATH_IMAGE005
for the first error estimation value
Figure 415776DEST_PATH_IMAGE001
And a second error estimate
Figure 455276DEST_PATH_IMAGE002
And correcting and adaptively realizing the online evaluation of the CT running error.
In a possible embodiment, the first error estimate is obtained in step 4
Figure 117201DEST_PATH_IMAGE001
And a second error estimate
Figure 787217DEST_PATH_IMAGE002
The process of making the correction includes:
when the group evaluation condition is met, the step 1 is adopted to obtain a first error evaluation value
Figure 370645DEST_PATH_IMAGE001
Using the error deviation value
Figure 417099DEST_PATH_IMAGE004
Correcting to obtain the error evaluation value of the CT to be evaluated
Figure 434995DEST_PATH_IMAGE018
。
When the group evaluation condition is not met, the step 2 is adopted to obtain a second error evaluation value
Figure 275912DEST_PATH_IMAGE002
Using the error deviation value
Figure 877795DEST_PATH_IMAGE005
Correcting to obtain the error evaluation value of the CT to be evaluated
Figure 196781DEST_PATH_IMAGE019
。
According to the self-adaptive current transformer error evaluation method provided by the embodiment of the invention, the thought of multidimensional feature extraction and integrated fusion learning is utilized, the physical information neural network algorithm after optimization of the chameleon algorithm is adopted, the complementation and mutual correction of evaluation results of small-scale groups (single group of CT) and large-scale groups (3 groups or more) are realized, and the accuracy and the adaptive range of CT error evaluation are improved.
It should be noted that, in the foregoing embodiments, the description of each embodiment has an emphasis, and reference may be made to the related description of other embodiments for a part that is not described in detail in a certain embodiment.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
While preferred embodiments of the present invention have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. Therefore, it is intended that the appended claims be interpreted as including preferred embodiments and all such alterations and modifications as fall within the scope of the invention.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.

Claims (8)

1. An adaptive current transformer error evaluation method, characterized in that the error evaluation method comprises:
step 1, taking each line CT on the same bus in the same transformer substation as an evaluation group S, constructing a physical relation between each CT individual error and a current vector sum of a measured value node according to a kirchhoff current law, obtaining a measurement error of each CT in the group by using a mathematical processing method, and recording the measurement error as a first error estimation value of the CT
Figure 230202DEST_PATH_IMAGE001
；
Step 2, establishing an individual CT error evaluation model by using an off-line modeling and on-line training mode; the input of the individual CT error evaluation model is a multi-dimensional characteristic parameter of CT, and the output is a second error estimation value of CT
Figure 457921DEST_PATH_IMAGE002
(ii) a Obtaining a second error estimation value of each CT to be evaluated in the evaluation group S based on the individual CT error evaluation model
Figure 555190DEST_PATH_IMAGE002
；
Step 3, adopting an improved physical information neural network method to estimate the first error value
Figure 388016DEST_PATH_IMAGE001
And said second error estimate
Figure 279749DEST_PATH_IMAGE002
Performing fusion to obtain a third error estimation value
Figure 401289DEST_PATH_IMAGE003
；
Step 4, respectively calculating first error estimated values
Figure 594592DEST_PATH_IMAGE001
The second error estimation value
Figure 16346DEST_PATH_IMAGE002
And a third error estimate
Figure 344559DEST_PATH_IMAGE003
Error deviation value therebetween
Figure 484553DEST_PATH_IMAGE004
And
Figure 189204DEST_PATH_IMAGE005
(ii) a Using said error deviation value
Figure 731044DEST_PATH_IMAGE004
And
Figure 230158DEST_PATH_IMAGE005
for the first error estimation value
Figure 358914DEST_PATH_IMAGE001
And said second error estimate
Figure 601676DEST_PATH_IMAGE002
And correcting and adaptively realizing the online evaluation of the CT running error.
2. The error estimation method according to claim 1, wherein the step 1 includes:
step 101, acquiring and screening current measurement values of stable sections of CT of each line under the same node of a transformer substation in real time, and constructing a monitoring data set;
102, constructing a relational expression between a current measurement value and an error of a target CT in the monitoring data set and current measurement values and errors of other CTs according to a kirchhoff current law;
step 103, calculating a current true value of the target CT by using the target phase current data and the rated transformation ratio of the target CT as a current reference value
Figure 998022DEST_PATH_IMAGE006
(ii) a Selecting three-phase current data of the CT of any line in the monitoring data set for state evaluation, and respectively taking the corresponding CT and phase as a target CT and a target phase when the current data is in a normal state;
step 104, taking the current measurement values of other CTs except the target CT in the monitoring data set as input, and taking the current reference value
Figure 668038DEST_PATH_IMAGE006
As output, training the LAPO-RBF neural network to obtain a LAPO-RBF neural network model;
105, obtaining the current reference value of each to-be-evaluated CT by using the LAPO-RBF neural network model, and calculating a first error estimation value of each to-be-evaluated CT according to the current reference value
Figure 782625DEST_PATH_IMAGE001
。
3. The error estimation method according to claim 2, characterized in that in step 102
The obtained relation is:
Figure 563499DEST_PATH_IMAGE007
(10)
Figure 79931DEST_PATH_IMAGE008
(11)
Figure 920848DEST_PATH_IMAGE009
、
Figure 24195DEST_PATH_IMAGE010
approximately 0,n is the number of lines, m representsA. A phase B or a phase C, and a phase C,
Figure 608761DEST_PATH_IMAGE011
is the real part of the current of the measured value,
Figure 979699DEST_PATH_IMAGE012
is the imaginary part of the current of the measured value,
Figure 991517DEST_PATH_IMAGE013
for each of the line current measurements,
Figure 815117DEST_PATH_IMAGE014
for the purpose of the corresponding ratio difference,
Figure 203373DEST_PATH_IMAGE015
the corresponding phase difference.
4. The error evaluation method according to claim 1, wherein the multidimensional feature parameter in step 2 comprises: working condition parameters, electromagnetic parameters and environmental parameters.
5. The error estimation method according to claim 1 or 4, characterized in that the step 2 comprises:
step 201, acquiring multi-dimensional characteristic parameters of each CT in an off-line detection process to construct an off-line data set; training an ANN model based on the offline data set to obtain a pre-training model;
step 202, acquiring multi-dimensional characteristic parameters of each CT in the online detection process, and constructing each segmented data set according to the current value range of the CT;
step 203, inputting each segmented data set into each pre-training model for training, and obtaining each individual CT error evaluation model corresponding to the current value range of each CT;
step 204, inputting the online monitoring data of the CT to be evaluated into the individual CT error evaluation model corresponding to the segment where the current value is positioned,the individual CT error evaluation model outputs a second error estimation value of the CT to be evaluated
Figure 163239DEST_PATH_IMAGE002
。
6. The error estimation method according to claim 1, wherein the step 3 adopts a method of improving a physical information neural network to obtain a fusion model of the error estimation value;
the input characteristic quantity of the fusion model is as follows:
Figure 345958DEST_PATH_IMAGE016
the output characteristic quantity is:
Figure 423898DEST_PATH_IMAGE017
；
Figure 350266DEST_PATH_IMAGE018
is the third error estimate for the CT to be evaluated.
7. The error evaluation method according to claim 1 or 6, wherein the construction process of the fusion model includes:
approximating a function through a fully-connected neural network, solving partial differential equation residual error and initial value residual error constraints by using an automatic differential technology, putting the partial differential equation residual error and initial value residual error constraints into a loss function as regular terms, and obtaining a neural network connection weight parameter W and a partial differential equation physical parameter b by using a chameleon algorithm.
8. The error estimation method of claim 1, wherein the first error estimate is evaluated in step 4
Figure 226955DEST_PATH_IMAGE001
And said second error estimate
Figure 846155DEST_PATH_IMAGE002
The process of making the correction includes:
when the group evaluation condition is met, obtaining the error evaluation value of the CT to be evaluated
Figure 644347DEST_PATH_IMAGE019
；
When the population evaluation condition is not met, the error evaluation value of the CT to be evaluated is obtained
Figure 374405DEST_PATH_IMAGE020
。
CN202211331032.0A 2022-10-28 2022-10-28 Self-adaptive current transformer error evaluation method Active CN115392141B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202211331032.0A CN115392141B (en) 2022-10-28 2022-10-28 Self-adaptive current transformer error evaluation method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202211331032.0A CN115392141B (en) 2022-10-28 2022-10-28 Self-adaptive current transformer error evaluation method

Publications (2)

Publication Number Publication Date
CN115392141A true CN115392141A (en) 2022-11-25
CN115392141B CN115392141B (en) 2023-03-24

Family

ID=84115027

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202211331032.0A Active CN115392141B (en) 2022-10-28 2022-10-28 Self-adaptive current transformer error evaluation method

Country Status (1)

Country Link
CN (1) CN115392141B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115878963A (en) * 2023-01-05 2023-03-31 国网江苏省电力有限公司营销服务中心 Capacitance voltage transformer metering error prediction method, system, terminal and medium
CN117151932B (en) * 2023-10-27 2024-01-12 武汉纺织大学 Method and system for predicting error state of non-stationary output current transformer

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1988306A (en) * 2006-11-18 2007-06-27 重庆市电力公司 Intelligent bus protecting method based on artificial nerve network model
CN111025222A (en) * 2019-12-31 2020-04-17 三峡大学 Electronic voltage transformer state monitoring method based on recursion principal component analysis
CN111537939A (en) * 2020-04-17 2020-08-14 武汉格蓝若智能技术有限公司 Voltage transformer state evaluation method and device based on multi-index fusion
US20200309829A1 (en) * 2019-04-01 2020-10-01 Dake He On-line monitoring system for the performance of the measurement equipment in the entire power grid based on wide-area synchronous measurement
CN113297797A (en) * 2021-06-02 2021-08-24 东南大学 XGboost-based electronic transformer measurement error state evaluation method and device
CN113702895A (en) * 2021-10-28 2021-11-26 华中科技大学 Online quantitative evaluation method for error state of voltage transformer
CN113935440A (en) * 2021-12-15 2022-01-14 武汉格蓝若智能技术有限公司 Iterative evaluation method and system for error state of voltage transformer
CN114692505A (en) * 2022-04-12 2022-07-01 广东电网有限责任公司 Error prediction method and device for capacitor voltage transformer
CN115166618A (en) * 2022-09-06 2022-10-11 武汉格蓝若智能技术有限公司 Current transformer error evaluation method for non-stable output

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1988306A (en) * 2006-11-18 2007-06-27 重庆市电力公司 Intelligent bus protecting method based on artificial nerve network model
US20200309829A1 (en) * 2019-04-01 2020-10-01 Dake He On-line monitoring system for the performance of the measurement equipment in the entire power grid based on wide-area synchronous measurement
CN111025222A (en) * 2019-12-31 2020-04-17 三峡大学 Electronic voltage transformer state monitoring method based on recursion principal component analysis
CN111537939A (en) * 2020-04-17 2020-08-14 武汉格蓝若智能技术有限公司 Voltage transformer state evaluation method and device based on multi-index fusion
CN113297797A (en) * 2021-06-02 2021-08-24 东南大学 XGboost-based electronic transformer measurement error state evaluation method and device
CN113702895A (en) * 2021-10-28 2021-11-26 华中科技大学 Online quantitative evaluation method for error state of voltage transformer
CN113935440A (en) * 2021-12-15 2022-01-14 武汉格蓝若智能技术有限公司 Iterative evaluation method and system for error state of voltage transformer
CN114692505A (en) * 2022-04-12 2022-07-01 广东电网有限责任公司 Error prediction method and device for capacitor voltage transformer
CN115166618A (en) * 2022-09-06 2022-10-11 武汉格蓝若智能技术有限公司 Current transformer error evaluation method for non-stable output

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
梁家碧: "基于扰动数据分析的并联电容器运行状态监测技术研究", 《中国优秀博士学位论文全文数据库（工程科技Ⅱ辑）》 *
胡琛等: "电子式电流互感器误差模型及误差状态预测方法", 《电力工程技术》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115878963A (en) * 2023-01-05 2023-03-31 国网江苏省电力有限公司营销服务中心 Capacitance voltage transformer metering error prediction method, system, terminal and medium
CN117151932B (en) * 2023-10-27 2024-01-12 武汉纺织大学 Method and system for predicting error state of non-stationary output current transformer

Also Published As

Publication number Publication date
CN115392141B (en) 2023-03-24

Similar Documents

Publication Publication Date Title
CN115392141B (en) Self-adaptive current transformer error evaluation method
CN107577870B (en) Power distribution network voltage power sensitivity robust estimation method based on synchronous phasor measurement
CN109165504B (en) Power system false data attack identification method based on anti-generation network
CN115480204A (en) Current transformer operation error online evaluation optimization method based on big data deduction
CN115469259A (en) RBF neural network-based CT error state online quantitative evaluation method and device
CN115508770B (en) KL-NB algorithm-based electric energy meter operation state online evaluation method
CN112651519A (en) Secondary equipment fault positioning method and system based on deep learning theory
Cramer et al. Bad data detection and handling in distribution grid state estimation using artificial neural networks
CN103324858A (en) Three-phase load flow state estimation method of power distribution network
Wang et al. A DRL-aided multi-layer stability model calibration platform considering multiple events
CN109066651A (en) The calculation method of wind-powered electricity generation-load scenarios limit transmitted power
CN106372440B (en) A kind of adaptive robust state estimation method of the power distribution network of parallel computation and device
CN112821420B (en) XGboost-based prediction method and system for dynamic damping factor and multidimensional frequency index in ASFR model
CN111371088A (en) Method and system for correcting SVG control strategy based on BP neural network
Mobeen et al. A Generalized Modeling and Analysis for Transformer Parameterization
CN113241736A (en) Method and system for line protection adaptability analysis of new energy sending
Andreoli et al. Topology validation via simultaneous state & topology estimation with phasor data processing capability
CN115954956A (en) Method and system for evaluating access capacity of distributed power supply of power distribution network
Rodríguez-García et al. Methodology for measurement-based load modeling considering integration of dynamic load models
CN112565180B (en) Power grid defense method, system, equipment and medium based on moving target defense
CN113806899A (en) Method and device for identifying topological relation of power distribution network and mobile terminal
CN113139295A (en) Method and system for estimating comprehensive state of power system
Yue et al. A power system harmonic problem based on the BP neural network learning algorithm
Zhu et al. Recursive Bayesian-based approach for online automatic identification of generalized electric load models in a multi-model framework
CN110796385A (en) Power system state estimation method, device, equipment and storage medium

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