CN110095744B - Error prediction method for electronic transformer - Google Patents
Error prediction method for electronic transformer Download PDFInfo
- Publication number
- CN110095744B CN110095744B CN201910270880.7A CN201910270880A CN110095744B CN 110095744 B CN110095744 B CN 110095744B CN 201910270880 A CN201910270880 A CN 201910270880A CN 110095744 B CN110095744 B CN 110095744B
- Authority
- CN
- China
- Prior art keywords
- electronic transformer
- sample data
- error
- data
- value
- 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 41
- 230000007613 environmental effect Effects 0.000 claims abstract description 45
- 238000012545 processing Methods 0.000 claims abstract description 32
- 238000012549 training Methods 0.000 claims abstract description 12
- 238000013528 artificial neural network Methods 0.000 claims description 35
- 230000006870 function Effects 0.000 claims description 24
- 238000009826 distribution Methods 0.000 claims description 19
- 238000013507 mapping Methods 0.000 claims description 10
- 210000002569 neuron Anatomy 0.000 claims description 10
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000010606 normalization Methods 0.000 claims description 6
- 210000002364 input neuron Anatomy 0.000 claims description 5
- 239000011159 matrix material Substances 0.000 claims description 5
- 210000004205 output neuron Anatomy 0.000 claims description 5
- 230000008569 process Effects 0.000 claims description 5
- 238000002759 z-score normalization Methods 0.000 claims description 2
- 230000002159 abnormal effect Effects 0.000 abstract description 8
- 238000011425 standardization method Methods 0.000 abstract description 4
- 238000010586 diagram Methods 0.000 description 9
- 238000004590 computer program Methods 0.000 description 7
- 238000012544 monitoring process Methods 0.000 description 7
- 230000005540 biological transmission Effects 0.000 description 6
- 239000013307 optical fiber Substances 0.000 description 5
- 238000003860 storage Methods 0.000 description 3
- 239000000126 substance Substances 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 230000007774 longterm Effects 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000008054 signal transmission Effects 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
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
Landscapes
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses an error prediction method for an electronic transformer, which comprises the following steps: collecting error data and environmental parameter data of the electronic transformer to generate sample data, and eliminating abnormal data; carrying out standardization processing on the sample data based on a Z-score standardization method; clustering historical data of the environmental parameters, and establishing an error prediction model of the electronic transformer through training and learning; and predicting the specific difference and the angular difference of the electronic transformer based on a prediction model according to the environmental parameter value. The advantages are that: according to the error estimation method, no physical model is required to be established, on the basis of a multi-dimensional data driving method, the online estimation of the error of the electronic transformer can be realized according to the error data and the environmental parameter data of the electronic transformer, the problem that the error and the environmental parameter of the electronic transformer do not have a determined functional relation is solved, and the error prediction accuracy of the electronic transformer is improved.
Description
Technical Field
The invention belongs to the field of power transmission and distribution equipment state evaluation and fault diagnosis, and particularly relates to an electronic transformer error prediction method based on a clustering neural network.
Background
The electronic transformer is one of key devices for digital, automatic, informatization and interaction of the intelligent transformer substation. After the electronic transformer is adopted, the problem of additional error in the transmission of the electromagnetic transformer to secondary equipment is fundamentally solved through the signal transmission and the optical fiber, and the accuracy of a measuring and metering system is greatly improved. However, from the viewpoint of field operation, the accuracy problem of the electronic transformer still occupies a large proportion. Although all in-operation electronic transformers pass type tests and delivery tests, the error yield of the transformers is generally low when the transformers are installed on site, and most transformers need to be subjected to error adjustment on site. Due to the change of external environment parameters, errors of the electronic transformer are changed, the errors of the electronic transformer are inconsistent in laboratory, field off-line operation and on-line operation, the stability of the error state of the transformer is poor, and the metering reliability of the electronic transformer is seriously influenced.
The error check of the electronic transformer generally adopts a method of regular maintenance or power failure maintenance, and the electronic transformer is checked by using a standard electromagnetic transformer, and the method comprises an offline checking method and an online checking method. These calibration methods cannot evaluate the long-term operating error of the electronic transformer, are difficult to implement on site, and require heavy labor operation.
The prior art includes an electronic transformer error prediction method based on a time sequence model, and a future change trend is predicted by determining the time correlation of historical state data. However, the method does not take the effect of external factors into account, and when the external environment changes greatly, a prediction result may have a large error.
The prior art also comprises an electronic transformer error prediction method based on an error back propagation neural network, wherein an air-core coil current transformer is equivalent to a multi-input single-output system, the degradation characteristic of the transformer error is approximated based on a feedforward neural network, and the specific difference of the transformer is predicted according to a neural network model. However, this method does not consider the influence of the abnormal values in the sample data on the structure and parameters of the model, and the model has disadvantages such as being prone to fall into a local minimum value and having a slow convergence rate.
In conclusion, the prior art cannot evaluate the long-term operation error of the electronic transformer, is difficult to implement on site, needs heavy labor operation, is greatly influenced by the external environment, and has the defects of large error, unreliability and the like of the on-site prediction result.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides an error prediction method for an electronic transformer.
To solve the above technical problems, the present invention provides
An error prediction method for an electronic transformer is characterized by comprising the following steps:
s1: collecting an error value and an environmental parameter value of the electronic transformer to generate sample data;
s2: carrying out standardization processing on sample data of the error value of the electronic transformer, and unifying the unit and magnitude of the error value of the electronic transformer and the unit and magnitude of the environmental parameter value;
clustering historical data of the environment parameters, and dividing the environment parameter data into a plurality of spaces with larger differences to obtain clustering centers of different classifications;
s3: training and learning the clustered error data and environment parameter data of the electronic transformer, and establishing a corresponding neural network to obtain a ratio difference prediction model and an angle difference prediction model of the electronic transformer;
s4: and predicting the ratio difference and the angle difference of the electronic transformer according to the environmental parameter values based on the ratio difference prediction model and the angle difference prediction model obtained by training.
Further, in step S1, the sample data distribution rule generated by the error of the electronic transformer and the environmental parameter value is uniformly expressed by β (g, h) distribution, where x is β (g, h), and x is ∈ [ a, b ], where x represents the sample data, a and b represent the minimum and maximum of the sample data, and g and h represent distribution parameters.
Further, in step S2, the sample data is normalized based on the Z-score normalization method, and the calculation formula is:
where x' represents the value of the normalization process, x represents the sample data,
represents the mean of the sample data, and σ represents the standard deviation of the sample data.
Further, in the step S2, K objects in the sample data are selected as clustering centers; calculating the Euclidean distance between the sample data and the clustering center:
where n represents the number of sample data, xiRepresenting the sample data value, UKRepresenting a cluster center, distributing sample data to a nearest center vector according to a nearest principle, and forming disjoint clusters by the data; taking the mean value of all sample data in different clusters as a new cluster center; and iteratively calculating the Euclidean distance between the sample data and the new clustering center until the new clustering center is equal to the original clustering center or the maximum iteration number is reached.
Further, in step S3, the ratio difference or the angle difference of the electronic transformer is used as an output variable, the environment parameter data is used as an input variable, and the output of the network is used as an output variableThe number of input neurons is p, the number of output neurons is q, and the number of implicit neurons and the number of objects are equal and are K; the ith input vector of the neural network is Xi＝(x1,x2,…,xp) Wherein x is1，x2，...xpAs an input vector XiOf the ith output vector is Yi＝(y1,y2,…,yq) Wherein y is1，y2，...yqAs an output vector YiThe output layer weight is W ═ ω1,ω2,…,ωK]TWherein ω is1，ω2，...ωKAs elements of the initial output layer weight vector W, the basis function center value is the cluster center [ U ]1,U2,…,UK]TWherein U is1,U2,…,UKT represents the transpose of the matrix as an element of the cluster center.
Further, the mapping function of the implicit neurons is a gaussian function:
and according to the weight of the output layer and the mapping function, solving the output of the neural network as follows:
ωjobtaining the final weight W of the output layer of the neural network according to the least square method for the elements of the weight vector W of the initial output layern＝[w1，w2，…，wk]TWherein w is1,w2,…,wKRepresents the final output layer weight vector WnOf (2) is used.
Further, the environmental parameter value Ci＝(c1,c2,…,cp) Substituting the input of the neural network established in the step S3, and predicting the ratio difference or angular difference of the electronic transformerThe values are:
using the error actual value f of the electronic transformeriFor reference, the prediction error is calculated:
wherein, c1,c2,…,cpElements representing environmental parameters, wjN is the number of compute points, which are the elements of the final output layer weight vector.
An error prediction system of an electronic transformer is characterized by comprising a data acquisition module, a standardization processing and clustering processing module, a prediction model establishing module and a prediction module which are sequentially connected;
the data acquisition module is used for acquiring an error value and an environmental parameter value of the electronic transformer to generate sample data;
the standardization processing and clustering processing module is used for standardizing sample data of an error value of the electronic transformer, unifying the unit and magnitude of the error value of the electronic transformer and the unit and magnitude of an environmental parameter value, clustering historical data of the environmental parameter, and dividing the environmental parameter data into a plurality of spaces with large differences to obtain clustering centers of different classifications;
the prediction model establishing module is used for training and learning the error data and the environmental parameter data of the electronic transformer after clustering, and establishing a corresponding neural network to obtain a specific difference prediction model and an angular difference prediction model of the electronic transformer;
and the prediction module predicts the specific difference and the angular difference of the electronic transformer according to the environmental parameter value based on the specific difference prediction model and the angular difference prediction model obtained by training.
Furthermore, the distribution rule of the sample data generated by the error of the electronic transformer and the environmental parameter value is uniformly expressed by adopting beta (g, h) distribution, x-beta (g, h), and x belongs to [ a, b ], wherein x represents the sample data, a and b represent the minimum value and the maximum value of the sample data, and g and h represent distribution parameters.
Go toStep one, the standardization processing and clustering processing module adopts a Z-score-based standardization method to standardize sample data, and the calculation formula is as follows:
where x' represents the value of the normalization process, x represents the sample data,
represents the mean of the sample data, σ represents the standard deviation of the sample data;
the standardization processing and clustering processing module selects K objects in the sample data as clustering centers; calculating the Euclidean distance between the sample data and the clustering center:
where n represents the number of sample data, xiRepresenting the sample data value, UKRepresenting a cluster center, distributing sample data to a nearest center vector according to a nearest principle, and forming disjoint clusters by the data; taking the mean value of all sample data in different clusters as a new cluster center; and iteratively calculating the Euclidean distance between the sample data and the new clustering center until the new clustering center is equal to the original clustering center or the maximum iteration number is reached.
Furthermore, the prediction model building module is used for taking the ratio difference or the angle difference of the electronic transformer as an output variable, taking the environment parameter data as an input variable, taking the number of input neurons of the network as p, taking the number of output neurons as q, and taking the number of implicit neurons and the number of objects which are equal to each other as K; the ith input vector of the neural network is Xi＝(x1,x2,…,xp) Wherein x is1，x2，...xpAs an input vector XiOf the ith output vector is Yi＝(y1,y2,…,yq) Wherein y is1，y2，...yqAs an output vector YiThe output layer weight is W ═ ω1,ω2,…,ωK]TWherein ω is1，ω2，...ωKAs elements of the initial output layer weight vector W, the basis function center value is the cluster center [ U ]1,U2,…,UK]TWherein U is1,U2,…,UKT represents the transpose of the matrix, an element of the cluster center;
the mapping function of the implicit neurons is a gaussian function:
Further, the prediction module predicts an environmental parameter value Ci＝(c1,c2,…,cp) Substituting the input of the neural network established in the step S4, and the predicted value of the ratio difference or the angular difference of the electronic transformer is:
using the error actual value f of the electronic transformeriFor reference, the prediction error is calculated:
wherein, c1,c2,…,cpElements representing environmental parameters, wjN is the number of compute points, which are the elements of the final output layer weight vector.
The invention achieves the following beneficial effects:
according to the error estimation method, no physical model is required to be established, on the basis of a multi-dimensional data driving method, the online estimation of the error of the electronic transformer can be realized according to the error data and the environmental parameter data of the electronic transformer, the problem that the error and the environmental parameter of the electronic transformer do not have a determined functional relation is solved, and the error prediction accuracy of the electronic transformer is improved.
Drawings
FIG. 1 is a schematic flow chart of an embodiment of the present invention;
FIG. 2 is a schematic diagram of an error state monitoring platform of the electronic transformer;
FIGS. 3(a), (b), (c), (d) and (f) are the results of clustering calculations of temperature, humidity, vibration, magnetic field and load parameters, respectively;
FIGS. 4(a) and (b) are neural network iteration results of a specific difference prediction model and an angular difference prediction model, respectively;
fig. 5(a) and (b) are a specific difference prediction result and an angular difference prediction result, respectively, of the electronic transformer according to the present invention;
the system comprises an air-core coil current transformer 1, an electromagnetic current transformer 2, an environment monitoring unit 3, an optical fiber remote transmission unit 4, a signal acquisition unit 5, a data processing unit 6, a time synchronization unit 7, a switch 8 and a server 9.
Detailed Description
The invention is further described below with reference to the accompanying drawings. The following examples are only for illustrating the technical solutions of the present invention more clearly, and the protection scope of the present invention is not limited thereby.
An electronic transformer error prediction method based on a clustering neural network comprises the following steps:
step 1: acquiring an error value and an environmental parameter value of the electronic transformer to generate a sample, and rejecting abnormal data in the sample based on beta (g, h) distribution;
the value range of the sample data capacity is 10000 to infinity, and preferably, the sample capacity can be selected to be 15000.
The distribution rule of sample data generated by the error of the electronic transformer and the environmental parameter value is uniformly expressed by adopting beta (g, h) distribution, namely x-beta (g, h), x belongs to [ a, b ], wherein x represents the sample data, a and b represent the minimum value and the maximum value of the sample data, and g and h represent distribution parameters.
Let u be (x-a)/(x-b), the estimates of distribution parameters g and h are:
wherein the content of the first and second substances,
the mean value of u is represented by,
represents the variance of u.
Estimation of sample data distribution range
Comprises the following steps:
wherein the content of the first and second substances,
σ is the standard deviation of the sample data, which is the mean of the sample data. If the sample data value is in
And
otherwise, the data is judged to be abnormal data and is removed.
Step 2: based on a Z-score standardization method, carrying out standardization processing on the sample data after the abnormal data are removed, and unifying the unit and magnitude of the error value and the environmental parameter value of the electronic transformer;
the method is characterized in that the sample data after the abnormal data are removed is subjected to standardization processing based on a Z-score standardization method, and is a methodThe difference between the number of fractions and the mean is divided by the standard deviation, and the calculation formula is as follows:
and step 3: clustering historical data of the environment parameters, and dividing the environment parameter data into a plurality of spaces with larger differences to obtain clustering centers of different classifications;
selecting K objects in the sample data as clustering centers, wherein the value range of K is 1-4, and preferably K is 4; calculating the Euclidean distance between the sample data and the clustering center:
where n represents the number of sample data, xiRepresenting the sample data value, UjRepresenting a cluster center, distributing sample data to a nearest center vector according to a nearest principle, and forming disjoint clusters by the data; taking the mean value of all sample data in the K categories as a new clustering center; and iteratively calculating the Euclidean distance between the sample data and the new clustering center until the new clustering center is equal to the original clustering center or the maximum iteration number is reached.
And 4, step 4: training and learning the clustered error data and environment parameter data of the electronic transformer, and establishing a corresponding neural network to obtain a ratio difference prediction model and an angle difference prediction model of the electronic transformer;
and establishing a radial basis function neural network by training and learning by taking the specific difference or the angular difference of the electronic transformer as an output variable and taking the environmental parameter data as an input variable. The number of input neurons of the network is p, the number of output neurons is q, the number of implicit neurons is K, the value range of p is 2-5, preferably, p is 5, and q is 1; the ith input vector of the neural network is Xi＝(x1,x2,…,xp) The ith output vector is Yi＝(y1,y2,…,yq) The output layer weight is W ═ ω1,ω2,…,ωK]TCenter value of basis functionAs a cluster center [ U1,U2,…Uj...,UK]T(ii) a The mapping function of the implicit neurons is a gaussian function:
wherein the content of the first and second substances,
according to the output layer weight and the mapping function, the output of the neural network can be obtained as follows:
obtaining the final weight W of the output layer of the neural network according to the least square methodn＝[w1，w2，…，wk]T……(7)。
And 5: and predicting the ratio difference and the angle difference of the electronic transformer according to the environmental parameter value based on the prediction model obtained by training.
Will depend on the environmental parameter value Ci＝(c1,c2,…,cp) Substituting the input of the neural network established in the step S4, and the predicted value of the ratio difference or the angular difference of the electronic transformer is:
using the error actual value f of the electronic transformeriFor reference, the prediction error is calculated:
where N is the number of calculation points.
The invention is further described with reference to the following figures and specific examples. The examples are illustrative and are intended to be illustrative of the invention and should not be construed as limiting the invention.
As shown in fig. 1, the error of the electronic transformer is predicted according to the following steps:
(1) and (3) establishing an error state monitoring platform of the electronic transformer shown in fig. 2, and acquiring error data and environmental parameter data of the electronic transformer. The platform includes: the system comprises an environment monitoring unit 3, an optical fiber remote transmission unit 4, a signal acquisition unit 5, a data processing unit 6 and a time synchronization unit 7. A0.2-level air-core coil current transformer 1 and a 0.2-level electromagnetic current transformer 2 are installed in the platform. The output of the electromagnetic current transformer 2 is used as a standard signal, and the comparison result of the error of the hollow coil current transformer 1 can be obtained. The environment monitoring unit 3 can collect environment parameters of the installation position of the mutual inductor, including parameters such as temperature, humidity, vibration, magnetic field and the like; the optical fiber remote transmission unit 4 standardizes the data of the environment monitoring unit and sends the data to a data processing unit 6; the data processing unit 6 transmits the data to the server 9 through the switch 8, and the monitoring data are stored in the server 9; the signal acquisition unit 5 acquires output data of the digitalized electromagnetic current transformer 2; the data processing unit 6 receives the output data of the signal acquisition unit 5 and the sampling value message data of the air-core coil current transformer 1 at the same time. Constructing an original random matrix D according to the environmental parameters and the error data of the hollow coil current transformer 1; the clock synchronization unit 7 constructs a synchronous clock system of the whole system and is responsible for synchronizing the optical fiber remote transmission unit 4, the data processing unit 6 and the signal acquisition unit 5. And (3) eliminating 2 groups of abnormal data in the data sample by using a formula (1) and a formula (2).
(2) And (4) carrying out standardization processing on the sample data after the abnormal data are removed by using a formula (3), and unifying the unit and the magnitude of the error value of the electronic transformer and the unit and the magnitude of the environmental parameter value.
(3) And (5) repeatedly iterating by using the formula (4) to obtain a clustering center of the environment parameter, wherein a clustering result is shown in figure 3.
(4) The specific difference or angular difference of the electronic transformer is used as an output variable, the temperature, the humidity, the vibration, the magnetic field and the load parameter are used as input variables, the weights of the output layers of the neural network are calculated by using the formulas (5) to (7), the clustering neural network is established, and the iteration result of the neural network is shown in fig. 4.
(5) The environment parameters are substituted into the neural network, the ratio difference or the angular difference of the electronic transformer is predicted by using the formula (8), the prediction result is shown in fig. 5, and it can be seen that the maximum absolute value of the ratio difference prediction error is 0.05%, and the maximum absolute value of the angular difference prediction error is 9.5'. By taking the error true value of the electronic transformer as a reference and calculating the prediction error by using the formula (9), the specific error prediction error can reach 7.4 percent, the angular difference prediction error can reach 9.5 percent, and the prediction result is consistent with the actual value. As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application 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 application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. 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 processor, 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. The above description is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, several modifications and variations can be made without departing from the technical principle of the present invention, and these modifications and variations should also be regarded as the protection scope of the present invention.
Claims (10)
1. An error prediction method for an electronic transformer is characterized by comprising the following steps:
s1: collecting an error value and an environmental parameter value of the electronic transformer to generate sample data;
s2: carrying out standardization processing on sample data of the error value of the electronic transformer, and unifying the unit and magnitude of the error value of the electronic transformer and the unit and magnitude of the environmental parameter value;
clustering historical data of the environment parameters, and dividing the environment parameter data into a plurality of spaces with larger differences to obtain clustering centers of different classifications;
s3: training and learning the clustered error data and environment parameter data of the electronic transformer, and establishing a corresponding neural network to obtain a ratio difference prediction model and an angle difference prediction model of the electronic transformer;
in step S3, the ratio difference or the angle difference of the electronic transformer is used as an output variable, the environment parameter data is used as an input variable, the number of input neurons of the network is p, the number of output neurons is q, and the number of implicit neurons and the number of objects are equal and are K; the ith input vector of the neural network is Xi＝(x1,x2,…,xp) Wherein x is1，x2，...xpAs an input vector XiOf the ith output vector is Yi＝(y1,y2,…,yq) Wherein y is1，y2，...yqAs an output vector YiThe output layer weight is W ═ ω1,ω2,…,ωK]TWherein ω is1，ω2，...ωKAs elements of the initial output layer weight vector W, the basis function center value is the cluster center [ U ]1,U2,…,UK]TWherein U is1,U2,…,UKT represents the transpose of the matrix, an element of the cluster center;
s4: and predicting the ratio difference and the angle difference of the electronic transformer according to the environmental parameter values based on the ratio difference prediction model and the angle difference prediction model obtained by training.
2. The method for predicting the error of the electronic transformer according to claim 1, wherein in step S1, the sample data distribution rules generated by the error of the electronic transformer and the environmental parameter values are uniformly expressed by β (g, h) distribution, x- β (g, h), x ∈ a, b ], where x represents the sample data, a and b represent the minimum and maximum values of the sample data, and g and h represent the distribution parameters.
3. The electronic transformer error prediction method of claim 1, wherein in step S2, the sample data is normalized based on a Z-score normalization method, and the calculation formula is:
where x' represents the value of the normalization process, x represents the sample data,
represents the mean of the sample data, and σ represents the standard deviation of the sample data.
4. The electronic transformer error prediction method according to claim 1, wherein in step S2, K objects in the sample data are selected as clustering centers; calculating the Euclidean distance between the sample data and the clustering center:
where n represents the number of sample data, xiRepresenting the sample data value, UKRepresenting a cluster center, distributing sample data to a nearest center vector according to a nearest principle, and forming disjoint clusters by the data; taking the mean value of all sample data in different clusters as a new cluster center; and iteratively calculating the Euclidean distance between the sample data and the new clustering center until the new clustering center is equal to the original clustering center or the maximum iteration number is reached.
5. The electronic transformer error prediction method of claim 1, wherein the mapping function of the implicit neurons is a gaussian function:
and according to the weight of the output layer and the mapping function, solving the output of the neural network as follows:
ωjobtaining the final weight W of the output layer of the neural network according to the least square method for the elements of the weight vector W of the initial output layern＝[w1，w2，…，wk]TWherein w is1,w2,…,wKRepresents the final output layer weight vector WnOf (2) is used.
6. The electronic transformer error prediction method of claim 5, wherein the environmental parameter value C isi＝(c1,c2,…,cp) Substituting the input of the neural network established in the step S3, and the predicted value of the ratio difference or the angular difference of the electronic transformer is:
using the error actual value f of the electronic transformeriFor reference, the prediction error is calculated:
wherein, c1,c2,…,cpElements representing environmental parameters, wjN is the number of compute points, which are the elements of the final output layer weight vector.
7. An error prediction system of an electronic transformer is characterized by comprising a data acquisition module, a standardization processing and clustering processing module, a prediction model establishing module and a prediction module which are sequentially connected;
the data acquisition module is used for acquiring an error value and an environmental parameter value of the electronic transformer to generate sample data;
the standardization processing and clustering processing module is used for standardizing sample data of an error value of the electronic transformer, unifying the unit and magnitude of the error value of the electronic transformer and the unit and magnitude of an environmental parameter value, clustering historical data of the environmental parameter, and dividing the environmental parameter data into a plurality of spaces with large differences to obtain clustering centers of different classifications;
the prediction model establishing module is used for training and learning the error data and the environmental parameter data of the electronic transformer after clustering, and establishing a corresponding neural network to obtain a specific difference prediction model and an angular difference prediction model of the electronic transformer;
the prediction model establishing module is used for taking the ratio difference or the angle difference of the electronic transformer as an output variable and taking the environment parameter data as an input variable, the number of input neurons of the network is p, the number of output neurons of the network is q, and the number of implicit neurons and the number of objects which are equal are K; the ith input vector of the neural network is Xi＝(x1,x2,…,xp) Wherein x is1，x2，...xpAs an input vector XiOf the ith output vector is Yi＝(y1,y2,…,yq) Wherein y is1，y2，...yqAs an output vector YiThe output layer weight is W ═ ω1,ω2,…,ωK]TWherein ω is1，ω2，...ωKAs elements of the initial output layer weight vector W, the basis function center value is the cluster center [ U ]1,U2,…,UK]TWherein U is1,U2,…,UKT represents the transpose of the matrix, an element of the cluster center;
the mapping function of the implicit neurons is a gaussian function:
and the prediction module predicts the specific difference and the angular difference of the electronic transformer according to the environmental parameter value based on the specific difference prediction model and the angular difference prediction model obtained by training.
8. The electronic transformer error prediction system of claim 7, wherein the sample data distribution rule generated by the electronic transformer error and the environmental parameter value is uniformly expressed by using β (g, h) distribution, x- β (g, h), x ∈ [ a, b ], where x represents sample data, a and b represent the minimum and maximum values of the sample data, and g and h represent distribution parameters.
9. The electronic transformer error prediction system of claim 7, wherein the normalization processing and clustering processing module normalizes the sample data by a Z-score-based normalization method, and the calculation formula is:
where x' represents the value of the normalization process, x represents the sample data,
represents the mean of the sample data, σ represents the standard deviation of the sample data;
the standardization processing and clustering processing module selects K objects in the sample data as clustering centers; calculating the Euclidean distance between the sample data and the clustering center:
where n represents the number of sample data, xiRepresenting the sample data value, UKRepresenting a cluster center, distributing sample data to a nearest center vector according to a nearest principle, and forming disjoint clusters by the data; taking the mean value of all sample data in different clusters as a new cluster center; and iteratively calculating the Euclidean distance between the sample data and the new clustering center until the new clustering center is equal to the original clustering center or the maximum iteration number is reached.
10. The electronic transformer error prediction system of claim 7, wherein the prediction module transforms the environmental parameter value Ci＝(c1,c2,…,cp) Substituting the input of the neural network established in the step S4, and the predicted value of the ratio difference or the angular difference of the electronic transformer is:
wjn is the number of compute points, which are the elements of the final output layer weight vector.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910270880.7A CN110095744B (en) | 2019-04-04 | 2019-04-04 | Error prediction method for electronic transformer |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910270880.7A CN110095744B (en) | 2019-04-04 | 2019-04-04 | Error prediction method for electronic transformer |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110095744A CN110095744A (en) | 2019-08-06 |
| CN110095744B true CN110095744B (en) | 2021-05-11 |
Family
ID=67444305
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910270880.7A Active CN110095744B (en) | 2019-04-04 | 2019-04-04 | Error prediction method for electronic transformer |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110095744B (en) |
Families Citing this family (27)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110632546B (en) * | 2019-09-25 | 2021-07-02 | 国网江苏省电力有限公司电力科学研究院 | Electronic transformer credibility evaluation method and device based on whole-network-domain evidence set |
| CN110632545B (en) * | 2019-09-25 | 2021-09-21 | 国网江苏省电力有限公司电力科学研究院 | Regression neural network-based error risk assessment method and device for electronic transformer |
| CN110542879B (en) * | 2019-09-30 | 2022-02-11 | 国网四川省电力公司电力科学研究院 | Method and system for predicting operation performance variation trend of capacitor voltage transformer |
| CN111160419B (en) * | 2019-12-13 | 2023-04-18 | 中国电力科学研究院有限公司 | Deep learning-based electronic transformer data classification prediction method and device |
| CN111537938B (en) * | 2020-03-31 | 2022-12-09 | 国网江西省电力有限公司电力科学研究院 | Error short-time prediction method for electronic transformer based on intelligent algorithm |
| CN111814390B (en) * | 2020-06-18 | 2023-07-28 | 三峡大学 | Voltage transformer error prediction method based on transfer entropy and wavelet neural network |
| CN111896905A (en) * | 2020-06-18 | 2020-11-06 | 中国电力科学研究院有限公司 | Power transformer operation error state identification method and device |
| CN112379322A (en) * | 2020-11-05 | 2021-02-19 | 广东电网有限责任公司佛山供电局 | On-line monitoring method for capacitor voltage transformer |
| CN112816754B (en) * | 2020-12-22 | 2022-07-12 | 深圳供电局有限公司 | Current compensation method and equipment for current transformer |
| CN112989587B (en) * | 2021-03-04 | 2022-04-26 | 国网湖南省电力有限公司 | Online analysis method and system for degradation cause of capacitive voltage transformer |
| CN112990598A (en) * | 2021-03-31 | 2021-06-18 | 浙江禹贡信息科技有限公司 | Reservoir water level time sequence prediction method and system |
| CN113218444A (en) * | 2021-03-31 | 2021-08-06 | 中国电力科学研究院有限公司 | Intelligent sensing method and device for operating environment of electric energy metering device |
| CN113156359B (en) * | 2021-04-16 | 2024-01-26 | 中国电力科学研究院有限公司 | Method and system for determining metering error of capacitive voltage transformer |
| CN113447879A (en) * | 2021-05-27 | 2021-09-28 | 国网山东省电力公司营销服务中心(计量中心) | Electric energy meter metering error prediction method, system and device based on multi-stress influence |
| CN113376562B (en) * | 2021-06-08 | 2022-08-09 | 国网重庆市电力公司营销服务中心 | CVT (continuously variable transmission) verification method, device and medium based on rolling time window-FCM (fuzzy c-means) clustering |
| CN113567904A (en) * | 2021-07-02 | 2021-10-29 | 中国电力科学研究院有限公司 | Method and system suitable for metering error of capacitive mutual inductor |
| CN113325357A (en) * | 2021-08-04 | 2021-08-31 | 华中科技大学 | Voltage transformer error evaluation method and system based on output time series difference |
| CN113743651A (en) * | 2021-08-04 | 2021-12-03 | 三峡大学 | Electronic voltage transformer error state prediction method based on machine learning and SVM algorithm |
| CN113702895B (en) * | 2021-10-28 | 2022-02-08 | 华中科技大学 | Online quantitative evaluation method for error state of voltage transformer |
| CN113777552B (en) * | 2021-11-09 | 2022-03-11 | 武汉格蓝若智能技术有限公司 | Method and system for predicting metering error of capacitive voltage transformer |
| CN113821938B (en) * | 2021-11-18 | 2022-02-18 | 武汉格蓝若智能技术有限公司 | Short-term prediction method and device for metering error state of mutual inductor |
| CN115480204A (en) * | 2022-09-29 | 2022-12-16 | 武汉格蓝若智能技术有限公司 | Current transformer operation error online evaluation optimization method based on big data deduction |
| CN115542230B (en) * | 2022-11-24 | 2023-03-24 | 武汉格蓝若智能技术股份有限公司 | Current transformer error estimation method and device based on diffusion model |
| CN115616470B (en) * | 2022-12-20 | 2023-09-22 | 国网福建省电力有限公司 | Method, system, equipment and medium for predicting metering error state of current transformer |
| CN115878963B (en) * | 2023-01-05 | 2023-06-23 | 国网江苏省电力有限公司营销服务中心 | Method, system, terminal and medium for predicting metering error of capacitive voltage transformer |
| CN116125361B (en) * | 2023-02-03 | 2023-07-18 | 华中科技大学 | Voltage transformer error evaluation method, system, electronic equipment and storage medium |
| CN118091527B (en) * | 2024-04-24 | 2024-07-09 | 威胜集团有限公司 | Voltage transformer error assessment method based on interpretive deep learning |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2914068A1 (en) * | 2014-02-28 | 2015-09-02 | Zhejiang Sungiant arts and crafts Co., Ltd. | Electronic transformer-controlled two-wire two-way LED (light-emitting diode) lamp string |
Family Cites Families (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP5858519B2 (en) * | 2011-08-31 | 2016-02-10 | 国立研究開発法人産業技術総合研究所 | Error evaluation system for electronic current transformers |
| CN102967842B (en) * | 2012-10-24 | 2014-11-05 | 重庆大学 | Method for on-line diagnosing gradually-changing fault of electronic current transformers |
| CN103412948B (en) * | 2013-08-27 | 2017-10-24 | 北京交通大学 | The Method of Commodity Recommendation and system of collaborative filtering based on cluster |
| CN103837770B (en) * | 2014-02-26 | 2016-05-11 | 广东电网有限责任公司电力科学研究院 | Power equipments defect detects maintaining method |
| CN104850859A (en) * | 2015-05-25 | 2015-08-19 | 电子科技大学 | Multi-scale analysis based image feature bag constructing method |
| CN106548021B (en) * | 2016-11-03 | 2018-01-05 | 华中科技大学 | A kind of electronic mutual inductor error state appraisal procedure based on higher-dimension random matrix |
| CN107817395B (en) * | 2017-09-05 | 2020-08-07 | 天津市电力科技发展有限公司 | Electricity stealing troubleshooting method based on power grid line operation error remote calibration |
| CN107972508A (en) * | 2017-11-27 | 2018-05-01 | 南京晓庄学院 | A kind of electric automobile charge power control method and control device |
| CN108801448A (en) * | 2018-07-09 | 2018-11-13 | 滨州职业学院 | A kind of urban traffic noise monitoring system of intelligent terminal monitoring |
| CN109307852A (en) * | 2018-09-06 | 2019-02-05 | 中国电力科学研究院有限公司 | A kind of method and system of the measurement error of determining electric automobile charging pile electric energy metering device |
-
2019
- 2019-04-04 CN CN201910270880.7A patent/CN110095744B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2914068A1 (en) * | 2014-02-28 | 2015-09-02 | Zhejiang Sungiant arts and crafts Co., Ltd. | Electronic transformer-controlled two-wire two-way LED (light-emitting diode) lamp string |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110095744A (en) | 2019-08-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110095744B (en) | Error prediction method for electronic transformer | |
| US11544426B2 (en) | Systems and methods for enhanced sequential power system model parameter estimation | |
| CN106951695B (en) | Method and system for calculating residual service life of mechanical equipment under multiple working conditions | |
| Lennox et al. | Process monitoring of an industrial fed‐batch fermentation | |
| US8155908B2 (en) | Systems and methods for real-time system monitoring and predictive analysis | |
| CN109818349B (en) | Power grid robust state prediction method based on multidimensional state matrix sliding matching | |
| CN111814390B (en) | Voltage transformer error prediction method based on transfer entropy and wavelet neural network | |
| RU2626780C1 (en) | Method and system of remote monitoring energy installations | |
| CN113239132B (en) | Online out-of-tolerance identification method for voltage transformer | |
| CN109491289A (en) | A kind of dynamic early-warning method and device for data center's dynamic environment monitoring | |
| CN109583520B (en) | State evaluation method of cloud model and genetic algorithm optimization support vector machine | |
| RU2649542C1 (en) | Method and system of remote monitoring of objects | |
| EP2316108A2 (en) | A method for predicting power usage effectiveness and data center infrastructure efficiency within a real-time monitoring system | |
| Ali | Time‐between‐events control charts for an exponentiated class of distributions of the renewal process | |
| CN106980910B (en) | Medium-and-long-term power load measuring and calculating system and method | |
| CN109298225B (en) | Automatic identification model system and method for abnormal state of voltage measurement data | |
| US20210232104A1 (en) | Method and system for identifying and forecasting the development of faults in equipment | |
| CN110968619A (en) | Hydraulic press self-learning fault diagnosis method and system based on Fault Tree (FTA) | |
| CN116663747B (en) | Intelligent early warning method and system based on data center infrastructure | |
| CN109598052B (en) | Intelligent ammeter life cycle prediction method and device based on correlation coefficient analysis | |
| Tao et al. | Abnormal detection of wind turbine based on SCADA data mining | |
| CN114978956A (en) | Method and device for detecting abnormal performance mutation points of network equipment in smart city | |
| CN116451514A (en) | Bridge structure state evaluation method and device and electronic equipment | |
| Wang et al. | A DRL-aided multi-layer stability model calibration platform considering multiple events | |
| CN115685045A (en) | Online evaluation method for voltage transformer |
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 |