CN114840591B - Determination method and device for sectional switch power data - Google Patents
Determination method and device for sectional switch power data Download PDFInfo
- Publication number
- CN114840591B CN114840591B CN202210451966.1A CN202210451966A CN114840591B CN 114840591 B CN114840591 B CN 114840591B CN 202210451966 A CN202210451966 A CN 202210451966A CN 114840591 B CN114840591 B CN 114840591B
- Authority
- CN
- China
- Prior art keywords
- data
- switch
- power data
- sectional switch
- sample
- 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 50
- 238000013528 artificial neural network Methods 0.000 claims abstract description 56
- 238000009826 distribution Methods 0.000 claims abstract description 42
- 238000012360 testing method Methods 0.000 claims description 81
- 238000012549 training Methods 0.000 claims description 41
- 230000000694 effects Effects 0.000 claims description 32
- 238000004422 calculation algorithm Methods 0.000 claims description 17
- 238000013499 data model Methods 0.000 claims description 16
- 238000012545 processing Methods 0.000 claims description 15
- 239000011159 matrix material Substances 0.000 claims description 14
- 238000012795 verification Methods 0.000 claims description 13
- 238000010606 normalization Methods 0.000 claims description 7
- 238000012897 Levenberg–Marquardt algorithm Methods 0.000 claims description 6
- 210000002569 neuron Anatomy 0.000 claims description 6
- 238000010276 construction Methods 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 abstract description 7
- 230000008569 process Effects 0.000 abstract description 7
- 238000010586 diagram Methods 0.000 description 10
- 238000007726 management method Methods 0.000 description 8
- 238000004590 computer program Methods 0.000 description 7
- 230000001537 neural effect Effects 0.000 description 5
- 230000006870 function Effects 0.000 description 4
- 238000013461 design Methods 0.000 description 3
- 238000003860 storage Methods 0.000 description 3
- 238000010200 validation analysis Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000007423 decrease Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/25—Integrating or interfacing systems involving database management systems
- G06F16/258—Data format conversion from or to a database
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/28—Databases characterised by their database models, e.g. relational or object models
- G06F16/284—Relational databases
- G06F16/285—Clustering or classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- 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
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Business, Economics & Management (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Economics (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The application relates to the technical field of power distribution network line loss calculation and discloses a method and a device for determining sectional switch power data. According to the application, the BP neural network is trained by using the current and power data of the sectional switch and the voltage data of the surrounding distribution transformer, so that the fitting of the power data of the sectional switch with only the current and the surrounding distribution voltage data is realized, the problem of power data missing in the process of calculating the line loss is solved, and the efficiency of line loss management is improved.
Description
Technical Field
The application relates to the technical field of power distribution network line loss calculation, in particular to a method and a device for determining sectional switch power data.
Background
In the current power distribution network operation economic index, the line loss rate is one of important assessment indexes for measuring the power distribution network planning design, production operation, operation management and enterprise economic benefit level. The accurate and simple line loss calculation and analysis method is beneficial to drawing out reasonable line loss reduction measures and examining the actual effects of the measures, and is also convenient to draw out proper line loss assessment indexes and plans, so that the line loss management work is guided and promoted.
Under the background, it is becoming more and more important to acquire more real and complete power system measurement data, because only reliable line loss calculation results can better formulate corresponding loss reduction strategies, and further, the power grid operation efficiency and the line loss comprehensive management level are improved. However, in the process of calculating the line loss of the power distribution network, the problem of power data loss on the sectional switch can occur, and the efficiency of comprehensive management of the line loss is seriously affected.
Disclosure of Invention
The application discloses a method and a device for determining power data of a sectionalized switch, which are used for solving the technical problem that the power data on the sectionalized switch is lost in the current calculation process of the line loss of a power distribution network, and the efficiency of the line loss comprehensive management is seriously affected.
The first aspect of the application discloses a method for determining sectional switch power data, which comprises the following steps:
acquiring a plurality of groups of specific sectional switch data, wherein the specific sectional switch is a sectional switch with power data, and any group of specific sectional switch data comprises specific sectional switch current data, specific sectional switch power data and distribution voltage data around the specific sectional switch;
formatting the data of the plurality of groups of specific sectional switches into a matrix form and dividing the matrix form into a practical sample and a test sample;
normalizing the practice sample and the test sample;
Using a BP neural network, taking standardized specific sectional switch current data and distribution transformer voltage data around a specific sectional switch in a standardized practical sample as the input of the BP neural network, taking corresponding specific sectional switch power data as the expected output of the BP neural network, establishing and training the neural network, and constructing an initial power data determination model;
Comparing and verifying the initial power data determining model with the standardized test sample, analyzing a data fitting effect, and determining a final power data model if the fitting effect meets a preset requirement;
And acquiring current data of the to-be-detected sectional switch and voltage distribution data around the to-be-detected sectional switch, and determining the power data of the to-be-detected sectional switch according to the final power data model, the current data of the to-be-detected sectional switch and the voltage distribution data around the to-be-detected sectional switch.
Optionally, the specific segment switch current data is specific segment switch single-phase current data or specific segment switch three-phase current data.
Optionally, the normalizing the practice sample and the test sample includes:
the practical sample and the test sample are normalized by means of a mean square error method, and the data sizes of the practical sample and the test sample are changed to [0,1].
Optionally, the normalizing the practice sample and the test sample using a mean variance method includes:
The practice sample and the test sample are normalized by the following formula:
Where x k represents a practical sample and a test sample, x min represents a minimum value of data in the practical sample and the test sample, and x max represents a maximum value of data in the practical sample and the test sample.
Optionally, comparing the initial power data determining model with the standardized test sample for verification, and analyzing the data fitting effect, further includes:
And if the fitting effect does not reach the preset requirement, the initial power data determining model is re-determined.
Optionally, the establishing and training of the neural network includes:
and (3) taking 70% of samples in the practical samples as training data, 15% of samples as verification data and 15% of samples as test data, and establishing and training the neural network.
Optionally, in the BP neural network, the number of hidden neurons is set to 10.
Optionally, the training algorithm of the BP neural network is a Levenberg-Marquardt algorithm, a Bayesian Regularization algorithm or a Scaled Conjugate Gradient algorithm.
Optionally, the fitting effect is determined by mean square error and goodness of fit.
The second aspect of the present application discloses a determining device for segment switching power data, where the determining device for segment switching power data is applied to the determining method for segment switching power data disclosed in the first aspect of the present application, and the determining device for segment switching power data includes:
the specific data acquisition module is used for acquiring a plurality of groups of specific sectional switch data, wherein the specific sectional switch is a sectional switch with power data, and any group of specific sectional switch data comprises specific sectional switch current data, specific sectional switch power data and distribution voltage data around the specific sectional switch;
The preprocessing module is used for formatting the data of the plurality of groups of specific sectionalizing switches into a matrix form and dividing the matrix form into a practice sample and a test sample;
a normalization processing module for normalizing the practice sample and the test sample;
The initial model building module is used for using the BP neural network, taking standardized specific sectional switch current data and specific sectional switch surrounding distribution transformer voltage data in a standardized practical sample as the input of the BP neural network, taking corresponding specific sectional switch power data as the expected output of the BP neural network, building and training the neural network, and building an initial power data determination model;
The final model construction module is used for comparing and verifying the initial power data determination model with the standardized test sample, analyzing the data fitting effect, and determining a final power data model if the fitting effect meets the preset requirement;
The application module is used for acquiring current data of the to-be-detected segmented switch and distribution voltage data around the to-be-detected segmented switch, and determining the power data of the to-be-detected segmented switch according to the final power data model, the current data of the to-be-detected segmented switch and the distribution voltage data around the to-be-detected segmented switch.
Optionally, the specific segment switch current data is specific segment switch single-phase current data or specific segment switch three-phase current data.
Optionally, the standardized processing module is configured to: the practical sample and the test sample are normalized by means of a mean square error method, and the data sizes of the practical sample and the test sample are changed to [0,1].
Optionally, the standardized processing module is configured to: the practice sample and the test sample are normalized by the following formula:
Where x k represents a practical sample and a test sample, x min represents a minimum value of data in the practical sample and the test sample, and x max represents a maximum value of data in the practical sample and the test sample.
Optionally, the final model building module is further configured to: and if the fitting effect does not reach the preset requirement, the initial power data determining model is re-determined.
Optionally, the initial model building module is configured to: and (3) taking 70% of samples in the practical samples as training data, 15% of samples as verification data and 15% of samples as test data, and establishing and training the neural network.
Optionally, in the BP neural network, the number of hidden neurons is set to 10.
Optionally, the training algorithm of the BP neural network is a Levenberg-Marquardt algorithm, a Bayesian Regularization algorithm or a Scaled Conjugate Gradient algorithm.
Optionally, the fitting effect is determined by mean square error and goodness of fit.
The application discloses a method and a device for determining sectional switch power data. In the method, a plurality of groups of specific sectional switch data are firstly obtained, a matrix form is formatted and then divided into a practical sample and a test sample, then the practical sample and the test sample are standardized, a BP neural network is utilized to build and train the neural network according to the standardized practical sample, an initial power data determining model is built, the initial power data determining model is further compared with the standardized test sample for verification, a final power data model is determined, and finally the power data of the sectional switch to be tested is determined through the final power data model. According to the application, the BP neural network is trained by using the current and power data of the sectional switch and the voltage data of the surrounding distribution transformer, so that the fitting of the power data of the sectional switch with only the current and the surrounding distribution voltage data is realized, the problem of power data missing in the process of calculating the line loss is solved, and the efficiency of line loss management is improved.
Drawings
In order to more clearly illustrate the technical solution of the present application, the drawings that are needed in the embodiments will be briefly described below, and it will be obvious to those skilled in the art that other drawings can be obtained from these drawings without inventive effort.
FIG. 1 is a schematic workflow diagram of a method for determining segment switching power data according to an embodiment of the present application;
FIG. 2 is a graph of variation of mean square error of an example sample, as disclosed in an embodiment of the present application;
FIG. 3 is an exemplary test sample goodness-of-fit graph disclosed in an embodiment of the present application;
FIG. 4 is a graph of the variation of mean square error of an example two-down sample disclosed in an embodiment of the present application;
FIG. 5 is a plot of goodness-of-fit of a test sample under example II disclosed in an embodiment of the present application;
Fig. 6 is a schematic structural diagram of a determining device for determining power data of a sectionalizing switch according to an embodiment of the present application.
Detailed Description
In order to solve the technical problem that the power data on the sectional switch is lost in the current calculation process of the line loss of the power distribution network and the efficiency of the comprehensive management of the line loss is seriously affected, the application discloses a method and a device for determining the power data of the sectional switch through the following embodiments.
The first embodiment of the present application discloses a method for determining segment switching power data, referring to a workflow diagram shown in fig. 1, the method for determining segment switching power data includes:
Step S1, a plurality of groups of specific sectional switch data are obtained, the specific sectional switch is a sectional switch with power data, and any group of specific sectional switch data comprises specific sectional switch current data, specific sectional switch power data and distribution voltage data around the specific sectional switch.
Further, the specific sectionalized switch current data is specific sectionalized switch single-phase current data or specific sectionalized switch three-phase current data.
Specifically, a specific sectional switch on a specific area line with power data is selected, and single-phase or three-phase current data, power data and surrounding distribution voltage data are collected.
And S2, formatting the data of the plurality of groups of specific sectionalizing switches into a matrix form and dividing the matrix form into a practice sample and a test sample.
Specifically, to match the programming calculations of Matlab, they are formatted in a matrix and multiple sets of specific segmented switch data are separated into practical and test samples.
And step S3, normalizing the practice sample and the test sample.
Further, the normalizing the practice sample and the test sample includes:
the practical sample and the test sample are normalized by means of a mean square error method, and the data sizes of the practical sample and the test sample are changed to [0,1].
Further, the normalizing the practice sample and the test sample using the mean variance method includes:
The normalization processing can ensure that the data sizes are relatively close, avoid calculating certain data with small sample number due to large size difference, and improve the data processing precision. After normalization, the data size of all samples became [0,1]. Currently, the mean variance method is mainly used for normalization, in particular for normalization of the practice sample and the test sample by the following formula:
Where x k represents a practical sample and a test sample, x min represents a minimum value of data in the practical sample and the test sample, and x max represents a maximum value of data in the practical sample and the test sample.
In particular, since the data sizes of different samples are typically different, direct input without processing will affect convergence and processing speed. Thus, the data samples are normalized by means of the mean variance method.
And S4, using the BP neural network, taking standardized specific sectional switch current data and distribution transformer voltage data around the specific sectional switch in the standardized practical sample as the input of the BP neural network, taking corresponding specific sectional switch power data as the expected output of the BP neural network, establishing and training the neural network, and constructing an initial power data determination model.
In some embodiments of the present application, the establishing and training of the neural network includes:
and (3) taking 70% of samples in the practical samples as training data, 15% of samples as verification data and 15% of samples as test data, and establishing and training the neural network.
In some embodiments of the present application, in the BP neural network, the number of hidden neurons is set to 10, and the specific number may be adjusted according to practical situations.
In some embodiments of the present application, the training algorithm of the BP neural network is a Levenberg-Marquardt algorithm, bayesian Regularization algorithm, or Scaled Conjugate Gradient algorithm.
Specifically, the establishment and training of the neural network is performed by calling a BP neural network toolbox in Matlab.
And S5, comparing and verifying the initial power data determining model with the standardized test sample, analyzing the data fitting effect, and determining a final power data model if the fitting effect meets the preset requirement. The preset requirements are predetermined according to actual application scenes.
Further, the comparing and verifying the initial power data determining model with the standardized test sample, analyzing the data fitting effect, and further includes:
And if the fitting effect does not reach the preset requirement, the initial power data determining model is re-determined.
In some embodiments of the application, the fitting effect is determined by means of a mean square error and a goodness of fit.
Specifically, the trained neural network is compared with the test sample for verification, and the data fitting effect is analyzed to facilitate verification of the training effect.
The BP neural network analysis is carried out by calling a BP neural network tool box in Matlab, and the specific steps are as follows: selecting Neural NET FITTING a tool APP; according to the data type possessed by the sectional switch, normalized three-phase current data and three-phase voltage data or single-phase current data and three-phase voltage data are selectively imported as input data, and normalized power data are imported as expected output data; the ratio of the sample size occupied by three types of data, namely training data, verification data and test data, is selected, wherein the ratio can be adjusted according to specific conditions, and the ratio is that 70% of samples are randomly selected as the training data, 15% as the verification data and 15% as the test data by default; determining the number of hidden neurons, typically 10, and then adjusting the specific number according to the practical situation; training algorithms are selected, wherein the training algorithms are generally divided into three types of Levenberg-Marquardt, bayesian Regularization and Scaled Conjugate Gradient, and default Levenberg-Marquardt algorithms are generally selected; according to the obtained result, the smaller the Mean Square Error (MSE) value is, the closer the fitting goodness (R) value is to 1, and the better the training effect is; if the obtained model cannot meet the preset requirement, repeating the steps until the required fitting precision is obtained.
And S6, acquiring current data of the to-be-detected sectional switch and distribution voltage data around the to-be-detected sectional switch, and determining the power data of the to-be-detected sectional switch according to the final power data model, the current data of the to-be-detected sectional switch and the distribution voltage data around the to-be-detected sectional switch.
According to the method for determining the sectional switch power data disclosed by the embodiment of the application, a plurality of groups of specific sectional switch data are firstly obtained, the specific sectional switch data are divided into a practice sample and a test sample after being formatted into a matrix form, then the practice sample and the test sample are standardized, a BP neural network is utilized to build and train the neural network according to the standardized practice sample, an initial power data determination model is built, the initial power data determination model is further compared with the standardized test sample for verification, a final power data model is determined, and finally the power data of the sectional switch to be tested is determined through the final power data model. According to the application, the BP neural network is trained by using the current and power data of the sectional switch and the voltage data of the surrounding distribution transformer, so that the fitting of the power data of the sectional switch with only the current and the surrounding distribution voltage data is realized, the problem of power data missing in the process of calculating the line loss is solved, and the efficiency of line loss management is improved.
It should be understood that, although the steps in the flowchart of fig. 1 are shown in sequence as indicated by the arrows, the steps are not necessarily performed in sequence as indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least a portion of the steps in fig. 1 may include a plurality of steps or stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of the steps or stages is not necessarily sequential, but may be performed in rotation or alternatively with at least a portion of the steps or stages in other steps or other steps.
The following is further described in connection with specific embodiments:
Example one: with the known three-phase current data of the sectionalizing switch and three-phase voltage data of the surrounding distribution transformer, power values of the corresponding sectionalizing switch need to be determined. The method for determining the segment switch power data disclosed by the embodiment comprises the following specific steps:
130 groups of data of the sectionalizer switch in a certain 10kV line are selected, and each group of data comprises a three-phase current value and a power value of the sectionalizer switch and a three-phase voltage value of surrounding distribution transformer.
The 130 sets of data were normalized according to the formula of the mean variance method.
According to the network structure design of the BP neural network, 6 input neural units (three-phase voltage and three-phase current data) and 1 output neural unit (active power data of the sectionalizing switch) are selected, and the iteration frequency of the network is 1000. 70% is selected as training data (90 groups), 15% as validation data (20 groups) and 15% as test data (20 groups). The results after training are shown in Table 1, and the MSE result is 4.4876e-5, the mean square error of the validation data.
TABLE 1
| Number of samples | Mean square error MSE | Goodness of fit R | |
| Training data | 90 | 2.15135e-5 | 9.99829e-1 |
| Validating data | 20 | 4.48756e-5 | 9.99815e-1 |
| Test data | 20 | 7.03161e-5 | 9.99702e-1 |
Fig. 2 shows in detail the variation of the average error of three sets of sample data during online training. It can be seen that during repeated online training, the average error of the samples gradually decreases, and the average error of the confirmed samples of the test samples gradually decreases, but improving generalization ability is the main purpose of online training. The goodness of fit R of the sample data is very close to 1, and the expected requirement is realized.
And then sampling data of the other section switch of the line in a certain time period is selected as a test sample of the trained model, the output result of the comparison model is compared with the actual power value, the result is shown in figure 3, and the fitting goodness R between the model fitting value and the target value is 0.96214, so that the effect is good.
Example two: the single-phase current data of the sectionalizing switch and the three-phase voltage data of the surrounding distribution transformer are known, and the power value of the corresponding sectionalizing switch needs to be determined. The method for determining the segment switch power data disclosed by the embodiment comprises the following specific steps:
130 groups of data of the sectionalizer in the 10kV line are selected, and each group of data comprises a single-phase current value and a power value of the sectionalizer and a three-phase voltage value of surrounding distribution transformer.
The 130 sets of data were normalized according to the formula of the mean variance method.
According to the network structure design of the BP neural network, 4 input neural units (three-phase voltage and single-phase current data) and 1 output neural unit (active power data of the sectionalizing switch) are selected, and the network iteration number is 1000. 70% is selected as training data (90 groups), 15% as validation data (20 groups) and 15% as test data (20 groups). The results after training are shown in Table 2, and the MSE result of the verification data is 5.42717e-4, which is larger than the value in scenario 1.
TABLE 2
| Number of samples | Mean square error MSE | Goodness of fit R | |
| Training data | 90 | 2.57490e-4 | 9.98061e-1 |
| Validating data | 20 | 5.42717e-4 | 9.96858e-1 |
| Test data | 20 | 6.60963e-4 | 9.96564e-1 |
Fig. 4 shows in detail the variation of the average error of three sets of sample data during online training.
Then, sampling data of another section switch of the line in a certain time period is selected as a test sample of the trained model, the output result of the comparison model is compared with the actual power value, the result is shown in fig. 5, the fitting goodness R between the model fitting value and the target value is 0.81166, the effect is inferior to the fitting goodness in the scene 1, but the total effect is in the error allowable range.
The following are examples of the apparatus of the present application that may be used to perform the method embodiments of the present application. For details not disclosed in the embodiments of the apparatus of the present application, please refer to the embodiments of the method of the present application.
The second embodiment of the present application discloses a determining device for segment switching power data, where the determining device for segment switching power data is applied to the determining method for segment switching power data disclosed in the first embodiment of the present application, referring to a schematic structural diagram shown in fig. 6, and the determining device for segment switching power data includes:
the specific data acquisition module 10 is configured to acquire multiple sets of specific segment switch data, where the specific segment switch is a segment switch with power data, and any one set of specific segment switch data includes specific segment switch current data, specific segment switch power data, and distribution voltage data around the specific segment switch.
The preprocessing module 20 is used for formatting the data of the plurality of groups of specific segment switches into a matrix form and dividing the data into practice samples and test samples.
A normalization processing module 30 for normalizing the practice sample and the test sample.
The initial model building module 40 is configured to use the BP neural network, take the normalized specific segment switch current data and the normalized specific segment switch surrounding distribution transformer voltage data in the normalized practical sample as an input of the BP neural network, and take the corresponding specific segment switch power data as an expected output of the BP neural network, perform building and training of the neural network, and build an initial power data determination model.
And the final model construction module 50 is configured to compare and verify the initial power data determination model with the standardized test sample, analyze the data fitting effect, and determine a final power data model if the fitting effect meets a preset requirement.
The application module 60 is configured to obtain current data of the to-be-detected sectionalizing switch and voltage distribution data around the to-be-detected sectionalizing switch, and determine power data of the to-be-detected sectionalizing switch according to the final power data model, the current data of the to-be-detected sectionalizing switch and the voltage distribution data around the to-be-detected sectionalizing switch.
Further, the specific sectionalized switch current data is specific sectionalized switch single-phase current data or specific sectionalized switch three-phase current data.
Further, the standardized processing module is used for: the practical sample and the test sample are normalized by means of a mean square error method, and the data sizes of the practical sample and the test sample are changed to [0,1].
Further, the standardized processing module is used for: the practice sample and the test sample are normalized by the following formula:
Where x k represents a practical sample and a test sample, x min represents a minimum value of data in the practical sample and the test sample, and x max represents a maximum value of data in the practical sample and the test sample.
Further, the final model building module is further configured to: and if the fitting effect does not reach the preset requirement, the initial power data determining model is re-determined.
Further, the initial model building module is configured to: and (3) taking 70% of samples in the practical samples as training data, 15% of samples as verification data and 15% of samples as test data, and establishing and training the neural network.
Further, in the BP neural network, the number of hidden neurons is set to 10.
Further, the training algorithm of the BP neural network is a Levenberg-Marquardt algorithm, a Bayesian Regularization algorithm or a Scaled Conjugate Gradient algorithm.
Further, the fitting effect is determined by means of a mean square error and a goodness of fit.
It will be appreciated by those skilled in the art that 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 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 flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations 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.
It should be finally understood that the foregoing embodiments are merely illustrative of the technical solutions of the present invention and not limiting the scope of protection thereof, and although the present invention has been described in detail with reference to the foregoing embodiments, it should be understood by those skilled in the art that various changes, modifications or equivalents may be made to the specific embodiments of the invention, and these changes, modifications or equivalents are within the scope of protection of the claims appended hereto.
Claims (10)
1. A method for determining segment switching power data, comprising:
acquiring a plurality of groups of specific sectional switch data, wherein the specific sectional switch is a sectional switch with power data, and any group of specific sectional switch data comprises specific sectional switch current data, specific sectional switch power data and distribution voltage data around the specific sectional switch;
formatting the data of the plurality of groups of specific sectional switches into a matrix form and dividing the matrix form into a practical sample and a test sample;
normalizing the practice sample and the test sample;
Using a BP neural network, taking standardized specific sectional switch current data and distribution transformer voltage data around a specific sectional switch in a standardized practical sample as the input of the BP neural network, taking corresponding specific sectional switch power data as the expected output of the BP neural network, establishing and training the neural network, and constructing an initial power data determination model;
Comparing and verifying the initial power data determining model with the standardized test sample, analyzing a data fitting effect, and determining a final power data model if the fitting effect meets a preset requirement;
And acquiring current data of the to-be-detected sectional switch and voltage distribution data around the to-be-detected sectional switch, and determining the power data of the to-be-detected sectional switch according to the final power data model, the current data of the to-be-detected sectional switch and the voltage distribution data around the to-be-detected sectional switch.
2. The method of claim 1, wherein the particular segment switch current data is a particular segment switch single-phase current data or a particular segment switch three-phase current data.
3. The method of determining segment switching power data of claim 1, wherein the normalizing the practice samples and the test samples comprises:
the practical sample and the test sample are normalized by means of a mean square error method, and the data sizes of the practical sample and the test sample are changed to [0,1].
4. A method of determining segment switching power data according to claim 3, wherein the normalizing the practice samples and the test samples using a mean-square error method comprises:
The practice sample and the test sample are normalized by the following formula:
Where x k represents a practical sample and a test sample, x min represents a minimum value of data in the practical sample and the test sample, and x max represents a maximum value of data in the practical sample and the test sample.
5. The method for determining the segment switching power data according to claim 1, wherein the comparing and verifying the initial power data determining model with the standardized test sample, analyzing the data fitting effect, further comprises:
And if the fitting effect does not reach the preset requirement, the initial power data determining model is re-determined.
6. The method for determining segment switching power data according to claim 1, wherein the establishing and training of the neural network comprises:
and (3) taking 70% of samples in the practical samples as training data, 15% of samples as verification data and 15% of samples as test data, and establishing and training the neural network.
7. The method of determining segment switching power data according to claim 1, wherein the number of hidden neurons in the BP neural network is set to 10.
8. The method of claim 1, wherein the training algorithm of the BP neural network is a Levenberg-Marquardt algorithm, bayesian Regularization algorithm, or Scaled Conjugate Gradient algorithm.
9. The method of claim 1, wherein the fitting effect is determined by means of a mean square error and a goodness of fit.
10. A determination apparatus of segment switching power data, characterized in that the determination apparatus of segment switching power data is applied to the determination method of segment switching power data according to any one of claims 1 to 9, the determination apparatus of segment switching power data comprising:
the specific data acquisition module is used for acquiring a plurality of groups of specific sectional switch data, wherein the specific sectional switch is a sectional switch with power data, and any group of specific sectional switch data comprises specific sectional switch current data, specific sectional switch power data and distribution voltage data around the specific sectional switch;
The preprocessing module is used for formatting the data of the plurality of groups of specific sectionalizing switches into a matrix form and dividing the matrix form into a practice sample and a test sample;
a normalization processing module for normalizing the practice sample and the test sample;
The initial model building module is used for using the BP neural network, taking standardized specific sectional switch current data and specific sectional switch surrounding distribution transformer voltage data in a standardized practical sample as the input of the BP neural network, taking corresponding specific sectional switch power data as the expected output of the BP neural network, building and training the neural network, and building an initial power data determination model;
The final model construction module is used for comparing and verifying the initial power data determination model with the standardized test sample, analyzing the data fitting effect, and determining a final power data model if the fitting effect meets the preset requirement;
The application module is used for acquiring current data of the to-be-detected segmented switch and distribution voltage data around the to-be-detected segmented switch, and determining the power data of the to-be-detected segmented switch according to the final power data model, the current data of the to-be-detected segmented switch and the distribution voltage data around the to-be-detected segmented switch.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210451966.1A CN114840591B (en) | 2022-04-24 | 2022-04-24 | Determination method and device for sectional switch power data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210451966.1A CN114840591B (en) | 2022-04-24 | 2022-04-24 | Determination method and device for sectional switch power data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114840591A CN114840591A (en) | 2022-08-02 |
| CN114840591B true CN114840591B (en) | 2024-06-07 |
Family
ID=82568666
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210451966.1A Active CN114840591B (en) | 2022-04-24 | 2022-04-24 | Determination method and device for sectional switch power data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114840591B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115808589A (en) * | 2023-02-06 | 2023-03-17 | 国网江苏省电力有限公司电力科学研究院 | Power distribution network time-sharing segmentation line loss abnormity diagnosis method and device |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6119112A (en) * | 1997-11-19 | 2000-09-12 | International Business Machines Corporation | Optimum cessation of training in neural networks |
| CN106229972A (en) * | 2016-08-16 | 2016-12-14 | 北京国能日新系统控制技术有限公司 | A kind of integrated based on many meteorological sources and the wind power forecasting method of segmentation modeling |
| CN106991531A (en) * | 2017-03-30 | 2017-07-28 | 国家电网公司 | Power distribution network project construction effect Post-assessment Method based on BP neural network |
| CN109657881A (en) * | 2019-01-14 | 2019-04-19 | 南京国电南自电网自动化有限公司 | A kind of neural network photovoltaic power generation prediction technique and system suitable for small sample |
| CN111523737A (en) * | 2020-05-29 | 2020-08-11 | 四川大学 | Automatic optimization-approaching adjusting method for operation mode of electric power system driven by deep Q network |
| CN112580174A (en) * | 2020-12-11 | 2021-03-30 | 国网辽宁省电力有限公司鞍山供电公司 | Power distribution network line loss rate calculation method based on genetic algorithm optimization neural network |
-
2022
- 2022-04-24 CN CN202210451966.1A patent/CN114840591B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6119112A (en) * | 1997-11-19 | 2000-09-12 | International Business Machines Corporation | Optimum cessation of training in neural networks |
| CN106229972A (en) * | 2016-08-16 | 2016-12-14 | 北京国能日新系统控制技术有限公司 | A kind of integrated based on many meteorological sources and the wind power forecasting method of segmentation modeling |
| CN106991531A (en) * | 2017-03-30 | 2017-07-28 | 国家电网公司 | Power distribution network project construction effect Post-assessment Method based on BP neural network |
| CN109657881A (en) * | 2019-01-14 | 2019-04-19 | 南京国电南自电网自动化有限公司 | A kind of neural network photovoltaic power generation prediction technique and system suitable for small sample |
| CN111523737A (en) * | 2020-05-29 | 2020-08-11 | 四川大学 | Automatic optimization-approaching adjusting method for operation mode of electric power system driven by deep Q network |
| CN112580174A (en) * | 2020-12-11 | 2021-03-30 | 国网辽宁省电力有限公司鞍山供电公司 | Power distribution network line loss rate calculation method based on genetic algorithm optimization neural network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114840591A (en) | 2022-08-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112488395B (en) | Method and system for predicting line loss of power distribution network | |
| CN109145957B (en) | Method and device for identifying and processing abnormal indexes of power distribution network based on big data | |
| CN111505433B (en) | Low-voltage transformer area indoor variable relation error correction and phase identification method | |
| CN113126019B (en) | Remote estimation method, system, terminal and storage medium for error of intelligent ammeter | |
| CN116125361B (en) | Voltage transformer error evaluation method, system, electronic equipment and storage medium | |
| CN109657613A (en) | Large scale electric network abnormal load recognition methods based on power method and parallel computing | |
| CN110726898B (en) | Power distribution network fault type identification method | |
| CN106058865A (en) | Power grid node voltage dip risk evaluation method | |
| CN114840591B (en) | Determination method and device for sectional switch power data | |
| CN109033513A (en) | Method for diagnosing fault of power transformer and diagnosing fault of power transformer device | |
| CN111612149A (en) | Main network line state detection method, system and medium based on decision tree | |
| CN115600933B (en) | Electric meter power quality detection method and system based on Internet of things | |
| CN110348540B (en) | Clustering-based method and device for screening transient power angle stability faults of power system | |
| CN115685045A (en) | Online evaluation method for voltage transformer | |
| CN114626487A (en) | Line-variable relation checking method based on random forest classification algorithm | |
| CN112528762B (en) | Harmonic source identification method based on data correlation analysis | |
| CN107274025B (en) | System and method for realizing intelligent identification and management of power consumption mode | |
| CN107257135B (en) | Method, device and equipment for judging unbalance treatment measures of power distribution area | |
| CN115879799A (en) | Transformer substation electric energy quality analysis method | |
| CN108595753B (en) | Method and device for optimizing electromagnetic transient recovery characteristic curve of fan | |
| CN109871998B (en) | Power distribution network line loss rate prediction method and device based on expert sample library | |
| CN114757291A (en) | Single-phase fault identification optimization method, system and equipment based on machine learning algorithm | |
| CN113627041A (en) | Time-varying dynamic load online modeling method and device based on multi-source data fusion | |
| CN113836125A (en) | Distribution transformer data cleaning and repairing method and system for distribution internet of things edge side | |
| CN112365024A (en) | High-voltage direct-current converter station energy efficiency prediction method and system based on deep learning |
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 |