CN108120900B - Power distribution network fault positioning method and system - Google Patents
Power distribution network fault positioning method and system Download PDFInfo
- Publication number
- CN108120900B CN108120900B CN201711403304.2A CN201711403304A CN108120900B CN 108120900 B CN108120900 B CN 108120900B CN 201711403304 A CN201711403304 A CN 201711403304A CN 108120900 B CN108120900 B CN 108120900B
- Authority
- CN
- China
- Prior art keywords
- power distribution
- distribution network
- layer
- network
- fault
- 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
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/08—Locating faults in cables, transmission lines, or networks
- G01R31/081—Locating faults in cables, transmission lines, or networks according to type of conductors
- G01R31/086—Locating faults in cables, transmission lines, or networks according to type of conductors in power transmission or distribution networks, i.e. with interconnected conductors
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Remote Monitoring And Control Of Power-Distribution Networks (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention discloses a power distribution network fault positioning method, which comprises the following steps: performing machine learning training on a deep neural network model framework comprising a plurality of layers of network modules and a bidirectional long-time memory network module, thereby obtaining an optimal deep neural network model; each monitoring terminal carries out wave recording on the working condition of the power distribution network to obtain wave recording data, and intercepts the wave recording data to obtain a fault waveform area; extracting the characteristics of the fault waveform region by using a multilayer network module in the optimal deep neural network model; each monitoring terminal uploads the characteristic data to a system main station, the system main station performs characteristic data aggregation, and the characteristic data of the monitoring terminals on the same transmission line are combined into a characteristic data sequence according to a power distribution network topological structure; and inputting the characteristic data sequence into the bidirectional long-time memory network module so as to obtain the relative position between each monitoring terminal and the fault point.
Description
Technical Field
The invention relates to the technical field of power detection, in particular to a power distribution network fault positioning method.
Background
CN103728532 discloses a method for locating a ground fault by installing a special distribution automation feeder terminal in a distribution line sectionalizing switch. According to the method, the distribution automation feeder terminals collect zero-sequence voltage 3U0 and zero-sequence current 3I0, then 3U0 and 3I0 are subjected to a series of processing and feature extraction, then the positions of switches where the current distribution automation feeder terminals are located relative to a ground fault point are judged by using preset fault judgment rules, and finally a fault section is located by combining a plurality of distribution automation feeder terminals.
According to the technical scheme, only fault information collected by a single terminal is used for judging when a fault is judged, comprehensive judgment of the fault information collected by a plurality of terminals is not effectively utilized, higher judgment accuracy can be achieved by comparing and judging characteristics among the plurality of terminals when a ground fault is utilized, the terminal only outputs semaphore information (0 or 1) on a fault path, the fault characteristic information is discarded after the fault judgment is completed, further analysis on the fault cannot be performed afterwards, meanwhile, updating of a preset fault judgment rule in the terminal can be completed only by updating a terminal program, a large amount of terminals need to be updated, and the terminals are in a work stop state for a period of time in the updating process.
CN104101812 discloses a fault detection and positioning method and system for a low-current grounding power distribution network. The system consists of a feeder line monitoring unit, a communication terminal and a system main station. And after the feeder line monitoring unit detects the suspected ground fault, the other two phases of transmission data are wirelessly and synchronously triggered. The system main station and the communication terminal adopt GPS time service, and the communication terminal and the feeder line monitoring unit carry out time service through a time division multiplexing wireless communication network. The system main station gathers the fault recording data of the three-phase feeder line monitoring unit of a plurality of points through the communication terminal, then extracts the transient signal of zero sequence voltage and zero sequence current from the recording in the main station, calculates the characteristic value, includes: amplitude, average value, differential value, integral value and combination thereof, transient zero sequence active power and zero sequence reactive power, calculating the similarity of transient zero sequence voltage and zero sequence current signal waveforms of each position, and preferentially judging each position on the screened suspected fault line according to the difference between the characteristic values and waveform similarity of the transient zero sequence voltage and zero sequence current before and after the ground fault point to locate the ground fault point.
However, in order to capture enough ground fault information, the fault recording will use a higher sampling rate, and usually a single recording will reach a size of tens of kbytes, thereby causing the following problems: 1. the recording transmission needs a long time, and has certain influence on the real-time performance of fault judgment; 2. in a season with high fault, multiple line faults may occur in a short time in the same main station region, and a large amount of data flow may rush into the main station in a period of time because collection of a single recording often passes through dozens of data interactions, so that the main station is required to have the capacity of processing a large amount of data flow for a long time, and the requirement on the performance of the main station is high; 3. most of the current power distribution terminals with fault recording adopt wireless networks and master station communication, a large amount of recording data can generate not small wireless charges, and the operation cost is high.
Meanwhile, in the fault location method in the prior art, the waveform feature extraction and fault location judgment are divided into two steps, namely, the multi-position original waveform in the power distribution network topology acquired by the terminal equipment in the power distribution network needs to be subjected to manual feature extraction firstly, and then fault location is carried out by using the features.
Disclosure of Invention
One of the technical problems to be solved by the invention is to provide fault location judgment of a power distribution network for an end-to-end neural network model, and simultaneously extract waveform characteristics to be built in a monitoring terminal of the power distribution network so as to reduce data interaction quantity between the monitoring terminal and a system main station.
In order to solve the technical problem, the invention provides a power distribution network fault positioning method, which comprises the following steps: performing machine learning training on a deep neural network model framework comprising a plurality of layers of network modules and a bidirectional long-time memory network module, thereby obtaining an optimal deep neural network model; each monitoring terminal carries out wave recording on the working condition of the power distribution network to obtain wave recording data, and intercepts the wave recording data to obtain a fault waveform area; extracting the characteristics of the fault waveform region by using a multilayer network module in the optimal deep neural network model; each monitoring terminal uploads the characteristic data to a system main station, the system main station performs characteristic data aggregation, and the characteristic data of the monitoring terminals on the same transmission line are combined into a characteristic data sequence according to a power distribution network topological structure; and inputting the characteristic data sequence into the bidirectional long-time memory network module so as to obtain the relative position between each monitoring terminal and the fault point.
In one embodiment, the multilayer network module is arranged in the monitoring terminal, and the monitoring terminal extracts the characteristics of the working condition recording.
In one embodiment, the multi-layer network module includes an input convolutional layer, a convolutional block, an average pooling layer, and a fully connected layer.
In one embodiment, the structure of the convolutional block may be a two-layer convolutional layer stack structure, or a multi-channel structure with each channel composed of two-layer convolutional layers stack, or a multi-channel structure with each channel containing 1 to 3 convolutional layers.
In one embodiment, residual connection is arranged between convolution blocks in the convolution layer area, and the residual connection refers to summing the input and the output of one convolution block and taking the sum result as input to the next convolution block.
In one embodiment, each long-short time memory unit in the bidirectional long-short time memory network corresponds to one monitoring terminal, and the sequence of the arrangement of the long-short time memory units corresponds to the arrangement mode of the feature data in the feature data sequence.
According to another aspect of the invention, a system for positioning the fault of the power distribution network is further provided, and the system comprises a system main station and a plurality of monitoring terminals arranged at different positions in the topology of the power distribution network; the system uses an end-to-end deep neural network to position and judge the fault of the power distribution network; the deep neural network comprises a multilayer network module and a bidirectional long-time and short-time memory network module, wherein the multilayer network module is arranged inside the monitoring terminal, and the bidirectional long-time and short-time memory network module is arranged inside the system main station.
In one embodiment, the multi-layer network module includes an input convolutional layer, a convolutional block, an average pooling layer, and a fully connected layer.
In one embodiment, the structure of the convolutional block may be a two-layer convolutional layer stack structure, or a multi-channel structure with each channel composed of two-layer convolutional layers stack, or a multi-channel structure with each channel containing 1 to 3 convolutional layers.
In one embodiment, each long-short time memory unit in the bidirectional long-short time memory network corresponds to one monitoring terminal, and the sequence of the arrangement of the long-short time memory units corresponds to the arrangement mode of the feature data in the feature data sequence.
The method for locating the fault of the power distribution network is further elaborated below.
As shown in fig. 1, the process diagram of the power distribution network fault location method of the present invention is that an end-to-end machine learning model is used for fault location determination, but in order to reduce the data transmission amount from a monitoring terminal to a system master station, the present invention integrates the extraction of waveform characteristics into the monitoring terminal. Specifically, fig. 2 is a schematic diagram of a deep neural network model framework for determining a fault location according to the present invention, which extracts waveform features using a multi-layer network module. It should be emphasized here that although the multi-layer network module is integrated in the monitoring terminal in the present invention, the monitoring terminal is far away from the system main station in physical location, but in terms of the deep neural network model for determining the fault location, the multi-layer network module is still a part of the neural network model. The multilayer network module is integrated with other parts of the neural network model positioned at the main station of the system, and the multilayer network module and the other parts of the neural network model jointly form the neural network model for judging the fault position.
When a fault occurs in the power distribution network, firstly, the monitoring terminal intercepts the wave recording data and intercepts a waveform section where the fault occurs. And then, performing feature extraction on the intercepted waveform section by using a multilayer network module built in the monitoring terminal. The structure of the multi-layer network module is shown in fig. 3, which includes an input convolutional layer, a convolutional block, an average pooling layer, and a full connection layer. Fig. 3a to 3c show specific structures of alternative convolution blocks in the present invention, wherein fig. 3a shows a two-layer convolution structure, which is formed by two layers of convolution layers being stacked. In FIG. 3b, a multi-channel structure is shown, and each channel is formed by two convolutional layers stacked together. Another multi-channel structure is shown in fig. 3c, each channel consisting of 1 to 3 convolutional layers. The convolution operation involved in the convolutional layer in the present invention adopts a convolution operation method known in the prior art, but the number of convolutional blocks, the specific structure of convolutional blocks, the length, width and number of convolution kernels of all convolutional layers, the number of layers of fully-connected layers and the number of neurons of fully-connected layers in the present invention, including but not limited to all the above parameters, are obtained by the hyper-parameter machine training of the present invention. In the present invention, a residual connection is added between the input and the output of a convolution block, that is, the sum of the input of each convolution block and the output of the convolution block is used as the output value of the convolution block, and then F (x) + x ═ H (x), where F (·) is the function of the convolution block, H (·) is the input of the next module, and x is the output of the previous module. And F (x) ═ h (x) — x, an increase in the residual x facilitates training of the F (·) parameter.
The feature data extracted by the multilayer network module is transmitted back to the master station, and because the feature data uploaded by each monitoring terminal are mutually independent and cannot reflect the interrelation among the monitoring terminals, the feature data need to be firstly collected in the master station of the system. The system master station classifies and collects all the characteristic data according to the topological structure of the power distribution network, collects the characteristic data uploaded by the power distribution network monitoring terminals under the same bus into a characteristic data sequence, as shown in a schematic diagram of the topological structure of the power distribution network shown in FIG. 4, it can be seen that four transmission lines are output from the transformer substation, when a fault occurs, all monitoring terminals upload the wave characteristic data, and the feature data classification performed by the system master station is to collect the feature data uploaded by all the monitoring terminals on the same line into a feature data sequence, for example, in the power distribution network topology structure shown in fig. 4, the feature data uploaded by the monitoring terminals a1, a2, a3 and a4 are collected into a feature data sequence, the arrangement mode of the characteristic data in the characteristic data sequence corresponds to the arrangement mode of the monitoring terminals in the topological structure of the power distribution network.
And then, inputting the characteristic data sequence into a bidirectional long-time and short-time memory network, wherein each long-time and short-time memory unit in the bidirectional long-time and short-time memory network corresponds to one monitoring terminal, and the arrangement sequence of the long-time and short-time memory units corresponds to the arrangement mode of the characteristic data in the characteristic data sequence, so that the topological information of the power distribution network can be integrated into the whole neural network model for fault location judgment.
The output of each long-short time memory unit in the bidirectional long-short time memory network passes through a multilayer full-connection layer and an output layer, and then the relative position relation of the monitoring terminal corresponding to the long-short time memory unit relative to the fault point can be obtained. For example, if the monitor terminal is located before the failure point, the relative positional relationship is 0, and if the monitor terminal is located after the failure point, the relative positional relationship is 1.
FIG. 5 is a schematic diagram of the hyper-parametric machine training process for the deep neural network model according to the present invention, wherein the purpose of the hyper-parametric machine training is to obtain all parameters required by the deep neural network model through training according to the training data set, the verification data set and the test data set, and to form an optimal hyper-parametric combination model of the deep neural network classifier. The machine training process is as follows:
a. inputting the deep neural network model frame into a hyper-parameter random generator;
b. forming a hyper-parameter combination model pool by a hyper-parameter random generator;
c. and testing each hyper-parameter combination model in the hyper-parameter combination model pool by using the test data set, finishing training if the test is passed, inputting the hyper-parameter combination model into the trained hyper-parameter combination model pool, optimizing the hyper-parameter combination model by using the training data set if the test is not passed, and testing again after the optimization until the model test is passed.
d. And verifying each hyper-parameter combination model in the trained hyper-parameter combination model pool by using a verification data set, wherein the hyper-parameter combination model passing the verification is the optimal hyper-parameter combination model.
The training data set, the verification data set and the test data set used in the super-parameter machine training process use about 35000 characteristic data sequences in total. The length of each characteristic data sequence is randomly distributed, that is, the number of the characteristic data corresponding to the monitoring terminal contained in each characteristic sequence is random. About 29400 training data sets were used and about 2800 testing and validation data sets were used, respectively, in the entire signature data sequence. The optimization method in the training process is batch Adam backward transmission, when the accuracy of the test data set is greater than 98% or the training exceeds 10000 rounds, the training is stopped, otherwise, the optimization is continued, and the combination with the highest accuracy of the verification data sets in the multiple hyper-parameter combination models is the optimal hyper-parameter combination model.
One or more embodiments of the present invention may have the following advantages over the prior art:
1. the method does not perform feature extraction on the waveform according to manually specified features, but the deep neural network model has the functions of feature extraction and fault point positioning judgment, and the identification accuracy can be further improved by the end-to-end training model.
2. In the invention, the modules for realizing the feature extraction in the deep neural network model are integrated in each monitoring terminal, so that the monitoring terminals upload waveform feature data to the system main station instead of working condition recording data, thereby obviously reducing the data interaction amount between each monitoring terminal and the system main station.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention and not to limit the invention. In the drawings:
FIG. 1 is a schematic flow chart of a power distribution network fault location method of the present invention;
FIG. 2 is a schematic diagram of a deep neural network model framework for fault location determination in accordance with the present invention
FIG. 3 is a schematic diagram of a multi-layer network module architecture of the present invention;
FIGS. 3a to 3c are schematic diagrams of convolution block structures according to the present invention;
FIG. 4 is a schematic diagram of a power distribution network topology;
FIG. 5 is a schematic diagram of the hyper-parametric machine training process for the deep neural network model according to the present invention.
FIG. 6 is a schematic diagram of an optimal deep neural network model according to an embodiment of the present invention;
fig. 7 is a schematic structural diagram of a multi-layer network module in the optimal deep neural network model according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to the accompanying drawings.
FIG. 6 is a schematic diagram of an optimal deep neural network model according to an embodiment of the present invention. The method is described below with reference to fig. 6.
First, machine training is performed on the deep neural network model framework schematic diagram for fault location determination of the present invention shown in fig. 2 according to the hyper-parametric machine training model shown in fig. 5, so as to obtain an optimal hyper-parametric model. In this embodiment, about 35000 feature data sequences are used in training of the training data set, the verification data set, and the test data set used in the super-parameter machine training process. The length of each characteristic data sequence is randomly distributed, that is, the number of the characteristic data corresponding to the monitoring terminal contained in each characteristic sequence is random. About 29400 training data sets were used and about 2800 testing and validation data sets were used, respectively, in the entire signature data sequence. The optimization method in the training process is batch Adam backward transmission, when the accuracy of the test data set is greater than 98% or the training exceeds 10000 rounds, the training is stopped, otherwise, the optimization is continued, and the combination with the highest accuracy of the verification data sets in the multiple hyper-parameter combination models is the optimal hyper-parameter combination model.
Fig. 7 is a schematic structural diagram of a multilayer network module in the optimal deep neural network model of this embodiment, where the specific structure of the optimal multilayer network module is as follows:
the width and length of the convolution kernel input to the convolution layer are 6 × 5, and the number is 8.
The convolution block i is a single-channel, two-layer convolution layer, where the width and length of the convolution kernel of the first convolution layer is 6 × 3, and the number is 8, and the width and length of the convolution kernel of the second convolution layer is 3 × 3, and the number is 16.
The convolution block II is set as a convolution layer with three channels, the channel a is a double-layer convolution layer, wherein the width and the length of convolution kernels of the first convolution layer are 1 multiplied by 5, the number of the convolution kernels is 16, the width and the length of convolution kernels of the second convolution layer are 1 multiplied by 5, and the number of the convolution kernels is 32. The channel b is a double layer convolutional layer, wherein the width and length of the convolutional kernel of the first convolutional layer is 1 × 5, the number is 16, the width and length of the convolutional kernel of the second convolutional layer is 1 × 5, and the number is 32. And the channel c is three convolutional layers, wherein the width and the length of a convolutional kernel of the first convolutional layer are 1 multiplied by 3, the number of the convolutional kernels is 16, the width and the length of a convolutional kernel of the second convolutional layer are 1 multiplied by 4, the number of the convolutional kernels is 16, the width and the length of a convolutional kernel of the third convolutional layer are 1 multiplied by 3, the number of the convolutional kernels is 32, and the sum of the results of the three channels of the convolutional block II is input into the convolutional block III.
The convolution block III is set as a convolution layer with three channels, the channel a is a double-layer convolution layer, wherein the width and the length of the convolution kernel of the first convolution layer are 1 multiplied by 2, the number of the convolution kernels is 32, the width and the length of the convolution kernel of the second convolution layer are 1 multiplied by 3, and the number of the convolution kernels is 64. The channel b is a double layer convolutional layer, wherein the width and length of the convolutional kernel of the first convolutional layer is 1 × 3, the number is 32, the width and length of the convolutional kernel of the second convolutional layer is 1 × 3, and the number is 64. And the channel c is three convolutional layers, wherein the width and the length of a convolutional kernel of the first convolutional layer are 1 multiplied by 3, the number of the convolutional kernels is 32, the width and the length of a convolutional kernel of the second convolutional layer are 1 multiplied by 3, the number of the convolutional kernels is 32, the width and the length of a convolutional kernel of the third convolutional layer are 1 multiplied by 3, the number of the convolutional kernels is 64, and the sum of the results of the three channels of the convolutional block III is input into the convolutional block IV.
In this embodiment, a residual connection is set between the convolution block i, the convolution block ii, and the convolution block iii, that is, the convolution block ii is input with the sum of the output result of the convolution block i and the output result of the convolution block ii, the convolution block iii is input with the sum of the output result of the convolution block i and the output result of the convolution block ii, and the sum of the output result of the convolution block ii and the output result of the convolution block iii is input to the average pooling layer. Parameter training of a convolution block I, a convolution block II and a convolution block III can be enhanced by setting residue connection. And the output of the average pooling layer enters two full-connection layers, the number of the neurons of the first full-connection layer is 24, and the number of the neurons of the second full-connection layer is 8.
And updating the multilayer network module with the structure into each monitoring terminal, and performing feature extraction on the waveform data by each monitoring terminal by using the multilayer network module. All feature data extracted by each monitoring terminal are uploaded to a system master station, the system master station collects all feature data according to a power distribution network topological structure to form feature data sequences representing different power distribution network transmission lines, each feature sequence is input into a bidirectional long-time and short-time memory network in an optimal deep neural network model shown in fig. 6, the number of features of each long-time and short-time memory unit in the long-time and short-time memory network is 8, the feature data output by the long-time and short-time memory units pass through two fully-connected layers, the number of neurons in the first fully-connected layer is 4, and the number of neurons in the second fully-connected layer is 1. And finally, outputting the fixed position between the monitoring terminal and the fault point by the output layer sigmoid.
The above description is only an embodiment of the present invention, and the protection scope of the present invention is not limited thereto, and any person skilled in the art should modify or replace the present invention within the technical specification of the present invention.
Claims (10)
1. A power distribution network fault positioning method is characterized by comprising the following steps:
performing machine learning training on a deep neural network model framework comprising a plurality of layers of network modules and a bidirectional long-time memory network module, thereby obtaining an optimal deep neural network model;
each monitoring terminal carries out working condition wave recording on the power distribution network to obtain wave recording data, and intercepts the wave recording data to obtain a fault waveform area;
carrying out feature extraction on the fault waveform region by utilizing a multilayer network module in the optimal deep neural network model to obtain feature data;
each monitoring terminal uploads the characteristic data to a system main station, the system main station performs characteristic data aggregation, and the characteristic data of the monitoring terminals on the same transmission line are combined into a characteristic data sequence according to the line position according to the topological structure of the power distribution network;
and inputting the characteristic data sequence into the bidirectional long-time memory network module so as to obtain the relative position between each monitoring terminal and the fault point.
2. The power distribution network fault location method of claim 1, wherein the multilayer network module is arranged inside a monitoring terminal, and the monitoring terminal extracts the characteristic of the working condition recording.
3. The method of claim 2, wherein the multi-layer network module comprises an input convolutional layer, a convolutional block, an average pooling layer, and a full connection layer.
4. The power distribution network fault location method according to claim 3, wherein the structure of the convolution block is a double layer convolution layer stacking structure, or is a multi-channel structure in which each channel is composed of double layer convolution layers, or is a multi-channel structure in which each channel includes 1 to 3 layers of convolution layers.
5. The power distribution network fault location method according to claim 4, wherein residual connections are arranged between convolution blocks in the convolution layer region, wherein the residual connections are that the input and the output of one convolution block are summed, and the summed result is transmitted as input to the next convolution block.
6. The method according to claim 1, wherein each of the two-way long-short-term memory cells in the bidirectional long-term memory network module corresponds to one monitoring terminal, and an arrangement sequence of the long-term memory cells corresponds to an arrangement manner of the feature data in the feature data sequence.
7. A system for power distribution network fault location, which uses the power distribution network fault location method of one of claims 1 to 6 for fault location, and comprises a system main station and a plurality of monitoring terminals arranged at different positions in a power distribution network topology; the system is characterized in that the system uses an end-to-end deep neural network to carry out positioning judgment on the faults of the power distribution network; the deep neural network comprises a multilayer network module and a bidirectional long-time and short-time memory network module, wherein the multilayer network module is arranged inside the monitoring terminal, and the bidirectional long-time and short-time memory network module is arranged inside the system main station.
8. The system for power distribution network fault location of claim 7, wherein the multi-layer network module comprises an input convolutional layer, a convolutional block, an average pooling layer, and a fully connected layer.
9. The system for power distribution network fault location according to claim 7, wherein each long-time memory unit in the bidirectional long-time memory network module corresponds to one monitoring terminal, and the sequence of the arrangement of the long-time memory units corresponds to the arrangement of the feature data in the feature data sequence.
10. A device for fault location of a power distribution network, characterized in that the device for fault location of a power distribution network uses the method for fault location of a power distribution network according to any of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711403304.2A CN108120900B (en) | 2017-12-22 | 2017-12-22 | Power distribution network fault positioning method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711403304.2A CN108120900B (en) | 2017-12-22 | 2017-12-22 | Power distribution network fault positioning method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108120900A CN108120900A (en) | 2018-06-05 |
| CN108120900B true CN108120900B (en) | 2020-02-11 |
Family
ID=62230984
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711403304.2A Active CN108120900B (en) | 2017-12-22 | 2017-12-22 | Power distribution network fault positioning method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108120900B (en) |
Families Citing this family (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109001593B (en) * | 2018-07-16 | 2020-09-25 | 北京映翰通网络技术股份有限公司 | Fault recording control method for power distribution network |
| CN110794255B (en) * | 2018-08-01 | 2022-01-18 | 北京映翰通网络技术股份有限公司 | Power distribution network fault prediction method and system |
| CN110794254B (en) * | 2018-08-01 | 2022-04-15 | 北京映翰通网络技术股份有限公司 | Power distribution network fault prediction method and system based on reinforcement learning |
| CN109389322A (en) * | 2018-10-30 | 2019-02-26 | 福州大学 | The disconnected broken lot recognition methods of grounded-line based on target detection and long memory models in short-term |
| CN109884475A (en) * | 2019-04-02 | 2019-06-14 | 云南电网有限责任公司大理供电局 | A kind of electric network fault detection method, device, system and storage medium |
| CN110261109B (en) * | 2019-04-28 | 2020-12-08 | 洛阳中科晶上智能装备科技有限公司 | Rolling bearing fault diagnosis method based on bidirectional memory cyclic neural network |
| CN110082640B (en) * | 2019-05-16 | 2021-11-30 | 国网安徽省电力有限公司 | Distribution network single-phase earth fault identification method based on long-time memory network |
| CN111999591B (en) * | 2019-05-27 | 2023-07-21 | 北京映翰通网络技术股份有限公司 | Method for identifying abnormal state of primary equipment of power distribution network |
| CN110456218B (en) * | 2019-08-08 | 2021-09-21 | 三峡大学 | Rapid fault line selection method based on power frequency increment coefficients before and after medium resistance switching |
| CN110969194B (en) * | 2019-11-21 | 2023-12-19 | 国网辽宁省电力有限公司电力科学研究院 | Cable early fault positioning method based on improved convolutional neural network |
| CN111273127B (en) * | 2020-01-22 | 2023-02-17 | 广东电网有限责任公司广州供电局 | Method and device for processing D-PMU fault wave recording data of power distribution network |
| CN111419220A (en) * | 2020-03-27 | 2020-07-17 | 山东省人工智能研究院 | Resonance sparse decomposition-based twelve-lead electrocardiosignal atrial fibrillation detection method |
| CN112051481B (en) * | 2020-08-12 | 2021-05-18 | 华中科技大学 | Alternating current-direct current hybrid power grid fault area diagnosis method and system based on LSTM |
| CN114662251B (en) * | 2020-12-24 | 2024-06-04 | 国网西藏电力有限公司拉萨供电公司 | Power distribution network fault positioning method based on deep neural network |
| CN116223971B (en) * | 2023-02-24 | 2023-12-12 | 合肥威标科技有限公司 | Auxiliary positioning system for fault area of power distribution network based on multi-source information |
| CN117368639A (en) * | 2023-10-18 | 2024-01-09 | 浙江德清迪生电力科技有限公司 | Power transmission line fault identification system and method based on artificial intelligence |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH04279870A (en) * | 1991-03-08 | 1992-10-05 | Hitachi Cable Ltd | Transmission line failure block location system |
| CN103091603A (en) * | 2013-01-14 | 2013-05-08 | 华北电力大学 | Breakdown intelligent classification and positioning method of electric transmission line |
| CN103454559A (en) * | 2013-09-02 | 2013-12-18 | 国家电网公司 | Power distribution network single-phase earth fault zone positioning method and positioning device |
| CN103713237A (en) * | 2013-12-30 | 2014-04-09 | 华北电力大学 | Power system electric transmission line fault diagnosis method |
| CN104834747A (en) * | 2015-05-25 | 2015-08-12 | 中国科学院自动化研究所 | Short text classification method based on convolution neutral network |
| CN105676067A (en) * | 2016-01-26 | 2016-06-15 | 云南电网有限责任公司电力科学研究院 | Fault diagnosis method and system of power transmission lines |
| CN106569089A (en) * | 2016-10-11 | 2017-04-19 | 国网上海市电力公司 | Power distribution line dynamic Information-based power distribution network fault monitoring method |
| CN106680576A (en) * | 2017-01-11 | 2017-05-17 | 福州大学 | Power distribution network internal overvoltage identification method based on partitioned time-frequency spectrum and deep learning algorithm |
| CN106980073A (en) * | 2017-05-23 | 2017-07-25 | 武汉大学 | A kind of two ends of electric transmission line fault recorder data matching process based on convolutional neural networks |
| CN107271925A (en) * | 2017-06-26 | 2017-10-20 | 湘潭大学 | The level converter Fault Locating Method of modularization five based on depth convolutional network |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105224984B (en) * | 2014-05-31 | 2018-03-13 | 华为技术有限公司 | A kind of data category recognition methods and device based on deep neural network |
-
2017
- 2017-12-22 CN CN201711403304.2A patent/CN108120900B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH04279870A (en) * | 1991-03-08 | 1992-10-05 | Hitachi Cable Ltd | Transmission line failure block location system |
| CN103091603A (en) * | 2013-01-14 | 2013-05-08 | 华北电力大学 | Breakdown intelligent classification and positioning method of electric transmission line |
| CN103454559A (en) * | 2013-09-02 | 2013-12-18 | 国家电网公司 | Power distribution network single-phase earth fault zone positioning method and positioning device |
| CN103713237A (en) * | 2013-12-30 | 2014-04-09 | 华北电力大学 | Power system electric transmission line fault diagnosis method |
| CN104834747A (en) * | 2015-05-25 | 2015-08-12 | 中国科学院自动化研究所 | Short text classification method based on convolution neutral network |
| CN105676067A (en) * | 2016-01-26 | 2016-06-15 | 云南电网有限责任公司电力科学研究院 | Fault diagnosis method and system of power transmission lines |
| CN106569089A (en) * | 2016-10-11 | 2017-04-19 | 国网上海市电力公司 | Power distribution line dynamic Information-based power distribution network fault monitoring method |
| CN106680576A (en) * | 2017-01-11 | 2017-05-17 | 福州大学 | Power distribution network internal overvoltage identification method based on partitioned time-frequency spectrum and deep learning algorithm |
| CN106980073A (en) * | 2017-05-23 | 2017-07-25 | 武汉大学 | A kind of two ends of electric transmission line fault recorder data matching process based on convolutional neural networks |
| CN107271925A (en) * | 2017-06-26 | 2017-10-20 | 湘潭大学 | The level converter Fault Locating Method of modularization five based on depth convolutional network |
Non-Patent Citations (2)
| Title |
|---|
| "Data-based line trip fault prediction in power systems using LSTM networks and SVM";SENLIN ZHANG 等;《IEEE Access》;20171221;第7675-7685页 * |
| "基于卷积神经网络的输电线路区内外故障判断及故障选相方法研究";魏东 等;《中国电机工程学报》;20160930;第36卷;第21-28页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108120900A (en) | 2018-06-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108120900B (en) | Power distribution network fault positioning method and system | |
| CN107609569B (en) | Power distribution network ground fault positioning method based on multi-dimensional feature vectors | |
| CN112041693B (en) | Power distribution network fault positioning system based on mixed wave recording | |
| CN110398663B (en) | Flexible direct current power grid fault identification method based on convolutional neural network | |
| CN107909118B (en) | Power distribution network working condition wave recording classification method based on deep neural network | |
| CN105512474B (en) | A kind of method for detecting abnormality of Transformer's Condition Monitoring data | |
| CN104502795B (en) | A kind of intelligent failure diagnosis method suitable for micro-capacitance sensor | |
| CN110829417B (en) | Electric power system transient stability prediction method based on LSTM double-structure model | |
| CN108154223B (en) | Power distribution network working condition wave recording classification method based on network topology and long time sequence information | |
| CN112149554B (en) | Training and fault detection methods of fault classification model and related devices | |
| CN106154180B (en) | Energy-storage battery charge/discharge anomaly detection method and detection system | |
| CN107340456A (en) | Power distribution network operating mode intelligent identification Method based on multiple features analysis | |
| CN102508076A (en) | Fault diagnosis device and method based on multi-agent system and wavelet analysis | |
| CN109444667A (en) | Power distribution network initial failure classification method and device based on convolutional neural networks | |
| CN111614066B (en) | Automatic setting method and system for relay protection setting value of power distribution network | |
| CN110794254B (en) | Power distribution network fault prediction method and system based on reinforcement learning | |
| CN105738765A (en) | Power distribution network fault positioning method based on feeder line terminals and genetic algorithm | |
| CN1805234A (en) | Pattern matching based security protection method for relay protection information of electric system in network environment | |
| CN113644748B (en) | System and method for simulating and monitoring substation equipment in area | |
| CN110807014B (en) | Cross validation based station data anomaly discrimination method and device | |
| CN114662251B (en) | Power distribution network fault positioning method based on deep neural network | |
| CN113406437B (en) | Power transmission line fault detection method for generating countermeasure network based on auxiliary classification | |
| CN115456109A (en) | Power grid fault element identification method and system, computer equipment and storage medium | |
| CN112465358A (en) | Voltage quality classification method and device based on support vector machine | |
| CN107016410A (en) | Power information acquisition system method for diagnosing faults and trouble-shooter |
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 |