CN108279364B - Power distribution network single-phase earth fault line selection method based on convolutional neural network - Google Patents
Power distribution network single-phase earth fault line selection method based on convolutional neural network Download PDFInfo
- Publication number
- CN108279364B CN108279364B CN201810088104.0A CN201810088104A CN108279364B CN 108279364 B CN108279364 B CN 108279364B CN 201810088104 A CN201810088104 A CN 201810088104A CN 108279364 B CN108279364 B CN 108279364B
- Authority
- CN
- China
- Prior art keywords
- convolutional neural
- neural network
- layer
- scale
- zero
- 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
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
- G01R31/088—Aspects of digital computing
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Supply And Distribution Of Alternating Current (AREA)
- Measurement Of Resistance Or Impedance (AREA)
Abstract
The invention relates to a power distribution network single-phase earth fault line selection method based on a convolutional neural network, which comprises the following steps: acquiring bus zero-sequence voltage and zero-sequence current signals of each feeder line; performing continuous wavelet transformation on each zero-sequence current signal according to a set decomposition scale; acquiring a time-scale wavelet coefficient gray scale image; and adopting a trained convolutional neural network algorithm to identify the fault feeder. The single-phase earth fault line selection method for the power distribution network based on the convolutional neural network can accurately identify fault feeders and non-fault feeders under various fault working conditions when single-phase earth faults occur.
Description
Technical Field
The invention relates to the field of power distribution networks, in particular to a single-phase earth fault line selection method for a power distribution network based on a convolutional neural network.
Background
With the improvement of national living standard and the continuous development of social and economic construction, people have increasingly increased electric power demand, and a power distribution system is used as an important part for connecting a power transmission system and users, and the safe and stable operation of the power distribution system has direct influence on the power utilization safety, the power utilization reliability and the power utilization benefits of the users. The probability of single-phase earth faults of the power distribution network reaches up to 80%, along with the complication of the structure of the power distribution network, system lines (including cable lines and cable-line mixed lines) in the power distribution network are gradually increased, so that the distributed capacitance of the system to the ground is increased, the capacitance current is also increased, and the insulation at the weak link of the system is easily damaged after long-time operation with faults, so that the faults are expanded to be two-point or multi-point earth short circuits; single-phase overvoltage caused by metallic grounding can burn a voltage transformer of a bus; arc grounding can also lead to over-voltages in the full network, so that faulty feeders must be found and removed accurately in time to prevent further propagation of the fault.
The core step of the single-phase earth fault feeder line identification research at home and abroad is the extraction of characteristic quantity. Due to the fact that different fault lines, different fault positions, different grounding resistances, different fault closing angles and other fault conditions can affect the size and the shape of the transient zero-sequence current, the accuracy of fault line selection is affected. A plurality of characteristic quantities are often required to be searched to represent the characteristic pattern of the single-phase earth fault signal, so that the identification purpose is achieved, and the classification algorithm applied to the identification problem of the single-phase earth fault feeder line of the power distribution network mainly adopts a relatively mature machine learning algorithm which does not have the self-learning capability. The deep learning algorithm can perform self-learning from a large amount of unmarked data, and the situations of overfitting or falling into a local optimal solution are prevented. The latest research results show that the deep learning algorithm improves the performance of classification and identification in multiple fields, and has a good application prospect in the classification and identification problem of the power system.
Disclosure of Invention
The invention aims to provide a power distribution network single-phase earth fault line selection method based on a convolutional neural network, so as to overcome the defects in the prior art.
In order to achieve the purpose, the technical scheme of the invention is as follows: a power distribution network single-phase earth fault line selection method based on a convolutional neural network is realized according to the following steps:
step S1: acquiring bus zero-sequence voltage and zero-sequence current signals of each feeder line;
step S2: performing continuous wavelet transformation on each zero-sequence current signal according to a preset decomposition scale;
step S3: acquiring a time-scale wavelet coefficient gray-scale image according to the time-frequency matrix acquired in the step S2;
step S4: and taking the acquired time-scale wavelet coefficient gray level images as training samples and test samples, and identifying the fault feeder line by adopting a trained convolutional neural network algorithm.
In an embodiment of the present invention, in step S1, a distribution network single-phase ground fault model is established, and a simulation waveform of bus zero sequence voltage from 1 cycle before a single-phase ground fault occurs to 2 power frequency cycles after the fault occurs and a simulation waveform of each feeder zero sequence current of 0.5 power frequency cycles after the fault occurs are intercepted by the distribution network single-phase ground fault model.
In an embodiment of the present invention, in step S2, the starting condition of the line selection process is set as whether the bus zero-sequence voltage is greater than a preset threshold, and the waveform decomposition is performed on the zero-sequence current signal of each feeder line by using a continuous wavelet transform decomposition method.
In one embodiment of the invention, the threshold value is 15% of the rated phase voltage.
In an embodiment of the present invention, in step S3, the number of sampling points of a zero-sequence current signal waveform of a feeder line is counted as m, and after a continuous wavelet transform with a decomposition scale of n, a time-scale wavelet coefficient matrix a is obtained as follows:
in an embodiment of the invention, db5 is selected as a wavelet basis function, the decomposition scale is 100, the number of sampling points is 100, and a grayscale image with an image sample size of 100 × 100 is obtained.
In an embodiment of the present invention, in the step S4, the convolutional neural network algorithm adopts a 12-layer convolutional neural network, which includes: 1 input layer, 5 convolutional layers, 4 pooling layers, 1 full-link layer and 1 output layer; the convolutional neural network algorithm is trained by using a back propagation algorithm.
In an embodiment of the present invention, the training of the convolutional neural network algorithm further includes the following steps:
step S41: initializing a convolutional neural network structure, comprising: the size and the number of convolution kernels of each convolution layer, the number of output feature maps, the size of a sampling window of each downsampling layer, a downsampling mode, a downsampling step length and a boundary continuation mode, and the number of samples of each batch of iteration and the number of iteration upper limits are set;
step S42: adjusting the image size of all time-scale wavelet coefficient gray scale image samples to be the size of an input image matched with an input layer, and dividing all samples into a training set and a testing set;
step S43: initializing parameters; initializing the weight omega and the bias item b of each layer into random numbers; initializing the hyper-parameters alpha and lambda into random numbers, and setting adjustment strategies of the two parameters when the training times are increased;
step S44: establishing a correlation matrix between layers;
step S45: starting to train the network, carrying out forward propagation once, calculating the activation value of each layer in sequence, and then calculating the error value between the actual output value of the output layer and the given type value;
step S46: performing back propagation on the error value obtained in the step S45, respectively calculating the weight of each layer and the adjustment quantity of the bias item, and judging whether the hyper-parameters need to be updated;
step S47: adjusting the weight and the bias item of each layer according to the adjustment quantity obtained in the step S46;
step S48: repeating the step S45 to the step S47 until the error meets the preset precision requirement or reaches the upper limit of the iteration times;
step S49: and after the training is finished, storing each updated latest parameter when the training is finished.
In an embodiment of the invention, the convolutional neural network algorithm is adopted to perform classification identification of the fault feeder line and the non-fault feeder line.
Compared with the prior art, the invention has the following beneficial effects:
1. the time-scale wavelet coefficient gray scale image is obtained by utilizing continuous wavelet transformation, and the time-frequency characteristics of the zero-sequence current signals can be completely extracted.
2. The method is combined with the convolutional neural network algorithm to identify the single-phase earth fault feeder line, extraction of a plurality of characteristic quantities is not needed, the convolutional neural network algorithm can independently learn the time-frequency characteristics of the single-phase earth fault zero-sequence current under various working conditions, namely the characteristics of a time-scale wavelet coefficient gray-scale map of each feeder line zero-sequence current signal, fault line selection can be carried out after training is completed, and the barrier that the identification characteristic quantities of the fault feeder line are difficult to extract is broken.
3. The method has strong adaptability, the measured waveform and the simulated waveform can obtain good effects, and the single-phase earth fault types of the power distribution network under the working conditions of different fault lines, different fault positions, different ground resistances, different fault closing angles, two-point earth, arc light earth, high resistance earth, noise interference, mutual inductor reverse connection, network structure change, distributed power supply access and the like can be accurately identified.
4. When the single-phase earth fault type is considered, different single-phase earth fault working conditions including a typical earth fault type, a special earth fault type, noise-resistant performance, network structure change, distributed power supply access and the like are fully considered, and good adaptability of the algorithm is fully displayed.
5. The method can directly identify zero sequence current simulation waveforms of all feeder lines with single-phase earth faults, also can directly identify zero sequence current waveforms of all feeder lines recorded by a field wave recording device, is tightly combined with engineering practice, and reliably reflects various conditions of single-phase earth faults in actual operation of a power system.
Drawings
Fig. 1 is a flowchart of a method for selecting a single-phase earth fault of a power distribution network based on a convolutional neural network according to an embodiment of the present invention.
Detailed Description
The embodiments of the present invention will be described more fully hereinafter with reference to the accompanying drawings, in which some, but not all embodiments of the invention are shown. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The invention provides a power distribution network single-phase earth fault line selection method based on a convolutional neural network, which comprises the following steps as shown in figure 1:
step S1: bus zero-sequence voltage and each feeder zero-sequence current signal;
step S2: performing continuous wavelet transformation on each zero-sequence current signal according to a set decomposition scale;
step S3: acquiring a time-scale wavelet coefficient gray scale image;
step S4: and adopting a trained convolutional neural network algorithm to identify the fault feeder.
Further, in step S1, a distribution network single-phase earth fault model is established, and the power distribution network single-phase earth fault model intercepts simulation waveforms of bus zero sequence voltage from 1 cycle before the single-phase earth fault occurs to 2 power frequency cycles after the single-phase earth fault occurs, and simulation waveforms of each feeder zero sequence current from 0.5 power frequency cycles after the single-phase earth fault occurs.
Further, wavelet transform is a nonlinear, non-stationary method of analyzing transient signals. The method has good localization property in time domain and frequency domain, so that the method is more accurate and reliable than Fourier transform and short-time Fourier transform, and the detection of singular and instantaneous fault signals becomes more accurate. The occurrence of the method leads us to carry out deeper analysis on the transient fault signals conditionally and accurately extract the characteristics which are beneficial to fault line selection, thereby improving the accuracy of fault line selection. The outstanding superiority of wavelet transform in processing nonlinear and non-stationary transient signal analysis is embodied as follows:
1) as long as the basis functions meeting the allowable wavelet conditions can be used as the wavelet basis functions, the flexibility of wavelet transformation is high;
2) the application of multi-resolution analysis makes the wavelet transformation process very fast;
3) the wavelet transform has the characteristic of time-frequency dual-domain property that signal characteristics can be revealed in a time-frequency domain.
In the present embodiment, based on the above advantages, by appropriately selecting the basic wavelet basis, the wavelet transform can be used to extract the transient fault feature, and the accuracy of fault line selection is improved.
Let psi (t) be E L2(R) is a square integrable function whose Fourier transform is psi (omega) when phi (omega) satisfies the condition
Then psi (t) is the mother wavelet function, and the translation and expansion transformation are carried out on psi (t) to obtain a group of two-dimensional bases called wavelet function cluster, and then the group of bases is usedTo represent or approximate a certain signal, i.e. the basic idea of wavelet transform. The selected mother wavelet function is psi (t), a scale factor a and a translation factor b are defined, and then the generated wavelet function cluster psia,b(t) is
For arbitrary functions x (t) at L2The integral continuous wavelet transform on (R) is defined as:
further, in step S2, using whether the bus zero-sequence voltage is greater than a preset threshold as a starting condition of the line selection process, and performing waveform decomposition on the zero-sequence current signal of each feeder line by using a continuous wavelet transform decomposition method. Preferably, the threshold is 15% of the nominal phase voltage.
Further, in step S3, the number of sampling points of the zero-sequence current signal waveform of a feeder line is counted as m, and a time-scale wavelet coefficient matrix a (a) is obtained after continuous wavelet transform with a decomposition scale of nij(i 1, 2.. n; j 1, 2.. m)) is:
preferably, db5 is selected as the wavelet basis function, the decomposition scale is 100, the number of sampling points is 100, and a grayscale image with the image sample size of 100 × 100 is obtained. The time-scale wavelet coefficient gray scale graph is used as the input of the convolutional neural network, the time-frequency characteristics of zero-sequence current signals of different feeder lines can be reflected, and the darker the gray scale color is, the larger the value of the representation coefficient is.
Further, in step S4, the convolutional neural network algorithm employs a 12-layer Convolutional Neural Network (CNN), including: 1 input layer, 5 convolutional layers, 4 pooling layers, 1 full-link layer and 1 output layer; training the convolutional neural network algorithm by using a Back Propagation (BP) algorithm, specifically comprising the steps of:
step S41: initializing a convolutional neural network structure, comprising: the size and the number of convolution kernels of each convolution layer, the number of output feature maps (convolution step default is 1), the size of a sampling window of each downsampling layer, a downsampling mode, a downsampling step and a boundary continuation mode and the like, and the number of iteration samples of each batch, the number of iteration upper limit times and the like are set;
step S42: adjusting the image size of all time-scale wavelet coefficient gray-scale image samples to be matched with the input image size required by an input layer, and dividing all samples into a training set and a testing set;
step S43: initializing parameters; initializing the weight omega and the bias term b of each layer to be random numbers close to 0; initializing the hyper-parameters alpha and lambda into random numbers which are small enough, and setting adjustment strategies of the two parameters when the training times are increased;
step S44: defining a correlation matrix between layers;
step S45: starting to train the network, carrying out forward propagation once, calculating the activation value of each layer in sequence, and then calculating the error value between the actual output value of the output layer and the given type value;
step S46: performing back propagation on the error value obtained in the step S45, respectively calculating the weight of each layer and the adjustment quantity of the bias item, and judging whether the hyper-parameters need to be updated;
step S47: adjusting the weight and the bias item of each layer according to the adjustment quantity obtained in the step S46;
step S48: repeating the step S45 to the step S47 until the error meets the preset precision requirement or reaches the upper limit of the iteration times;
step S49: and after the training is finished, storing each updated latest parameter when the training is finished.
In order to make those skilled in the art further understand the technical solution proposed by the present invention, the following description is made with reference to specific embodiments.
In the embodiment, a power distribution network single-phase earth fault model or a wave recording device built by simulation software is used for acquiring the waveform of bus zero-sequence voltage and each feeder zero-sequence current, and a preset bus zero-sequence voltage and feeder zero-sequence current waveform interval is intercepted; when the bus zero-sequence voltage exceeds a set threshold value (15% of rated phase voltage), a line selection process is started, then, the zero-sequence current signals of the feeder line are decomposed by using continuous wavelet transformation, a time-frequency matrix is obtained by selecting a proper decomposition scale, and a time-scale wavelet coefficient gray-scale image is obtained and is used as a training sample and a test sample of a convolutional neural network, so that the effective identification of the single-phase ground fault feeder line is realized.
Each feeder zero sequence current signal selected in the embodiment is derived from PSCAD/EMTDC software to build a 10kV power distribution network model for acquiring zero sequence current data. Both training and test samples were 9320.
The identification steps of the single-phase earth fault feeder line are as follows:
(1) acquisition of time-scale wavelet coefficient gray scale map
According to the technical scheme provided by the invention, the simulation waveform of the zero sequence current of each feeder line of 0.5 power frequency cycle (100 sampling points) after the single-phase earth fault occurs is intercepted. And (3) respectively carrying out continuous wavelet transformation processing on the transient zero-sequence current waveform of each feeder line in the system, selecting db5 as a wavelet basis function, and obtaining a time-scale wavelet coefficient gray-scale image with the image sample size of 100 multiplied by 100, wherein the decomposition scale is 100.
(2) Single phase ground fault feeder type identification
According to the technical scheme provided by the invention, the input image of the input layer is a wavelet coefficient gray scale image of 100 multiplied by 100;
the convolution layer 1 adopts 32 convolution kernels with 5 × 5 to perform convolution operation on an input image, the convolution moving step length is 1, and 32 characteristic graphs with 96 × 96 are output;
the down-sampling layer 1 adopts a maximum value down-sampling mode, the size of a sampling window is 3 multiplied by 3, the horizontal and longitudinal step lengths are both 2, the boundary continuation mode adopts the supplement of the upper side and the left side, and 32 characteristic graphs of 48 multiplied by 48 are output;
the convolution layer 2 adopts a convolution kernel of 1024 5 × 5 to check 32 characteristic graphs for convolution operation, the convolution moving step length is 1, and the boundary continuation mode adopts four sides of upper, lower, left and right to supplement to obtain 32 output characteristic graphs of 48 × 48;
the down-sampling layer 2 adopts an average value down-sampling mode, the size of a sampling window is 3 multiplied by 3, the horizontal and longitudinal step lengths are both 2, the boundary continuation mode adopts the supplement of the upper side and the left side, and 32 characteristic graphs of 24 multiplied by 24 are output;
the convolution layer 3 adopts 32 × 64 to 2048 convolution kernels with 5 × 5 to perform convolution operation on 32 feature maps, the convolution moving step length is 1, and the boundary continuation mode adopts four-side supplement including upper, lower, left and right to obtain 64 24 × 24 output feature maps;
the down-sampling layer 3 adopts an average value down-sampling mode, the size of a sampling window is 3 multiplied by 3, the horizontal and longitudinal step lengths are both 2, the boundary continuation mode adopts the supplement of the upper side and the left side, and 64 characteristic graphs of 12 multiplied by 12 are output;
the convolution layer 4 performs convolution operation on 64 feature maps by adopting 64 × 64-4096 convolution kernels of 5 × 5, the convolution moving step is 1, and 64 output feature maps of 8 × 8 are obtained;
the down-sampling layer 4 adopts an average value down-sampling mode, the size of a sampling window is 3 multiplied by 3, the horizontal and longitudinal step lengths are both 2, the boundary continuation mode adopts the supplement of the upper side and the left side, and 64 feature graphs of 4 multiplied by 4 are output;
the convolution layer 5 performs convolution operation on 64 feature maps by adopting 8192 convolution kernels with the size of 64 × 128 being 8192 and the size of 5 × 5, wherein the convolution moving step is 1, and 128 output feature maps with the size of 1 × 1 are obtained;
the full connection layer expands the 128 characteristic graphs output by the down sampling layer according to columns, and the 128 characteristic graphs are stacked to form a 128 multiplied by 1 characteristic vector and are fully connected with the output layer;
the output layer outputs a 2 x 1 type discrimination vector.
Each element value in the output result is [0,255], the position number of the maximum element is taken as the finally judged fault feeder type number, and the corresponding relation is 1: a faulty feeder line; -1: a non-faulty feeder; the number of samples for each batch of iteration is set to be 40, and the upper limit number of iterations is set to be 8000.
And (4) identification result: the recognition accuracy reaches more than 98.26%.
The above are preferred embodiments of the present invention, and all changes made according to the technical scheme of the present invention that produce functional effects do not exceed the scope of the technical scheme of the present invention belong to the protection scope of the present invention.
Claims (6)
1. A power distribution network single-phase earth fault line selection method based on a convolutional neural network is characterized by comprising the following steps:
step S1: acquiring bus zero-sequence voltage and zero-sequence current signals of each feeder line;
step S2: performing continuous wavelet transformation on each zero-sequence current signal according to a preset decomposition scale;
step S3: acquiring a time-scale wavelet coefficient gray-scale image according to the time-frequency matrix acquired in the step S2;
step S4: taking the acquired time-scale wavelet coefficient gray-scale image as a training sample and a test sample, and identifying a fault feeder line by adopting a trained convolutional neural network algorithm; in step S1, a distribution network single-phase earth fault model is established, and a simulation waveform of bus zero-sequence voltage of 2 power frequency cycles from 1 cycle before a single-phase earth fault occurs to 1 cycle after the fault occurs and a simulation waveform of each feeder zero-sequence current of 0.5 power frequency cycles after the fault occurs are intercepted by the distribution network single-phase earth fault model; in step S2, taking whether the bus zero-sequence voltage is greater than a preset threshold as a starting condition of the line selection process, and performing waveform decomposition on the zero-sequence current signal of each feeder line by using a continuous wavelet transform decomposition method.
2. The convolutional neural network-based single-phase ground fault line selection method for the power distribution network of claim 1, wherein the threshold is 15% of the rated phase voltage.
3. The method for selecting the single-phase earth fault of the power distribution network based on the convolutional neural network as claimed in claim 1, wherein in the step S3, the number of sampling points of the zero-sequence current signal waveform of the feeder line is counted asmOn a decomposed scale ofnObtaining a time-scale wavelet coefficient matrix after the continuous wavelet transformationAComprises the following steps:
4. the single-phase earth fault line selection method for the power distribution network based on the convolutional neural network as claimed in claim 3, wherein db5 is selected as a wavelet basis function, the decomposition scale is 100, the number of sampling points is 100, and a gray-scale map with the image sample size of 100 x 100 is obtained.
5. The method for selecting the single-phase earth fault of the power distribution network based on the convolutional neural network as claimed in claim 1, wherein in the step S4, the convolutional neural network algorithm adopts a 12-layer convolutional neural network, which comprises: 1 input layer, 5 convolutional layers, 4 pooling layers, 1 full-link layer and 1 output layer; the convolutional neural network algorithm is trained by using a back propagation algorithm.
6. The convolutional neural network based single-phase ground fault line selection method for the power distribution network of claim 5, wherein the training of the convolutional neural network algorithm further comprises the following steps:
step S41: initializing a convolutional neural network structure, comprising: the size and the number of convolution kernels of each convolution layer, the number of output feature maps, the size of a sampling window of each downsampling layer, a downsampling mode, a downsampling step length and a boundary continuation mode, and the number of samples of each batch of iteration and the number of iteration upper limits are set;
step S42: adjusting the image size of all time-scale wavelet coefficient gray scale image samples to be matched with the input image size of an input layer, and dividing all samples into a training set and a test set;
step S43: initializing parameters; initializing each layer of weight omega and bias item b into random numbers; initializing hyper-parameters alpha and lambda into random numbers, and setting adjustment strategies of the two parameters as the training times increase;
step S44: establishing a correlation matrix between layers;
step S45: starting to train the network, carrying out forward propagation once, calculating the activation value of each layer in sequence, and then calculating the error value between the actual output value of the output layer and the given type value;
step S46: performing back propagation on the error value obtained in the step S45, respectively calculating the weight of each layer and the adjustment quantity of the bias item, and judging whether the hyper-parameters need to be updated;
step S47: adjusting the weight and the bias item of each layer according to the adjustment quantity obtained in the step S46;
step S48, repeating the step S45 ~ and the step S47 until the error meets the preset precision requirement or reaches the upper limit of the iteration times;
step S49: and after the training is finished, storing each updated latest parameter when the training is finished.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810088104.0A CN108279364B (en) | 2018-01-30 | 2018-01-30 | Power distribution network single-phase earth fault line selection method based on convolutional neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810088104.0A CN108279364B (en) | 2018-01-30 | 2018-01-30 | Power distribution network single-phase earth fault line selection method based on convolutional neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108279364A CN108279364A (en) | 2018-07-13 |
| CN108279364B true CN108279364B (en) | 2020-01-14 |
Family
ID=62805775
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810088104.0A Active CN108279364B (en) | 2018-01-30 | 2018-01-30 | Power distribution network single-phase earth fault line selection method based on convolutional neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108279364B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113945862A (en) * | 2021-10-18 | 2022-01-18 | 广东电网有限责任公司东莞供电局 | Method, device and equipment for identifying high-resistance grounding fault of power distribution network |
Families Citing this family (29)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109241944A (en) * | 2018-09-07 | 2019-01-18 | 福州大学 | A kind of distribution network failure recognition methods based on improvement multi-category support vector machines |
| CN109324266B (en) * | 2018-11-21 | 2021-06-22 | 国网电力科学研究院武汉南瑞有限责任公司 | Distribution network ground fault analysis method based on deep learning |
| CN109617045A (en) * | 2018-11-21 | 2019-04-12 | 宁波恒晨电力建设有限公司 | Reduce the power distribution network method for protecting and device of low tension switch trip risk |
| CN109444667B (en) * | 2018-12-17 | 2021-02-19 | 国网山东省电力公司电力科学研究院 | Power distribution network early fault classification method and device based on convolutional neural network |
| CN109782126B (en) * | 2018-12-27 | 2021-01-01 | 上海交通大学 | Power distribution network early fault detection method based on humanoid concept learning |
| CN109613402B (en) * | 2019-02-14 | 2020-11-27 | 福州大学 | Power distribution network high-resistance grounding fault detection method based on wavelet transformation and neural network |
| CN110333088B (en) * | 2019-04-19 | 2020-09-29 | 北京化工大学 | Caking detection method, system, device and medium |
| CN110007198B (en) * | 2019-05-16 | 2021-03-02 | 福州大学 | Single-phase earth fault starting method |
| CN112240965A (en) * | 2019-07-17 | 2021-01-19 | 南京南瑞继保工程技术有限公司 | Grounding line selection device and method based on deep learning algorithm |
| CN110346692A (en) * | 2019-08-20 | 2019-10-18 | 国网河南省电力公司电力科学研究院 | A kind of wire selection method for power distribution network single phase earthing failure based on time-frequency image information |
| CN110618353B (en) * | 2019-10-22 | 2021-04-06 | 北京航空航天大学 | Direct current arc fault detection method based on wavelet transformation + CNN |
| CN111404130B (en) * | 2020-04-14 | 2022-03-22 | 国网内蒙古东部电力有限公司赤峰供电公司 | Novel power distribution network fault detection method and fault self-healing system based on quick switch |
| CN111914732A (en) * | 2020-07-28 | 2020-11-10 | 北京交通大学 | Insulation fault diagnosis method, device, equipment and computer readable storage medium |
| CN111679158A (en) * | 2020-08-04 | 2020-09-18 | 国网内蒙古东部电力有限公司呼伦贝尔供电公司 | Power distribution network fault identification method based on synchronous measurement data similarity |
| CN114076855A (en) * | 2020-08-10 | 2022-02-22 | 施耐德电器工业公司 | Inrush current detection method and apparatus for transformer, and computer-readable storage medium |
| CN114076902A (en) * | 2020-08-10 | 2022-02-22 | 施耐德电器工业公司 | Ground fault directivity detection method, device and computer readable storage medium |
| CN112114231B (en) * | 2020-09-18 | 2023-10-10 | 广西大学 | CNN fault line selection method with continuous learning capability |
| CN112180210B (en) * | 2020-09-24 | 2021-08-10 | 华中科技大学 | Power distribution network single-phase earth fault line selection method and system |
| CN112180217A (en) * | 2020-09-30 | 2021-01-05 | 福州大学 | Power distribution network single-phase earth fault section positioning method and system |
| CN112946425A (en) * | 2021-02-08 | 2021-06-11 | 长沙理工大学 | Fault positioning method for mining travelling wave time-frequency domain characteristics by utilizing deep learning |
| DE102021103121A1 (en) | 2021-02-10 | 2022-08-11 | Infineon Technologies Ag | Methods and devices for detecting an arc |
| CN113092934B (en) * | 2021-03-23 | 2022-05-13 | 武汉大学 | Single-phase earth fault judgment method and system based on clustering and LSTM |
| CN113285431A (en) * | 2021-06-10 | 2021-08-20 | 云南电网有限责任公司迪庆供电局 | Intelligent power distribution network information physical fusion protection system and method |
| CN114252739B (en) * | 2021-12-24 | 2023-11-03 | 国家电网有限公司 | Power distribution network single-phase earth fault discrimination method, system, equipment and storage medium |
| CN115047296B (en) * | 2022-08-15 | 2022-10-25 | 四川轻化工大学 | Power distribution network fault section positioning method |
| CN115684835B (en) * | 2022-10-28 | 2023-05-23 | 贵州电网有限责任公司信息中心 | Power distribution network fault diagnosis method, device, equipment and storage medium |
| CN117314883B (en) * | 2023-10-27 | 2024-04-16 | 国网四川省电力公司电力科学研究院 | Power distribution network fault line selection method and system based on EWT and VGGNet |
| CN117590158B (en) * | 2024-01-18 | 2024-04-19 | 华中科技大学 | Abnormal state identification method, device and control system of power distribution network |
| CN117909909A (en) * | 2024-03-19 | 2024-04-19 | 青岛鼎信通讯股份有限公司 | Arc grounding fault identification method for power distribution network |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101924354A (en) * | 2010-04-19 | 2010-12-22 | 昆明理工大学 | Artificially neural network routing method for distribution network failure by using S-transforming energy sampling property |
| CN103163430A (en) * | 2013-03-29 | 2013-06-19 | 昆明理工大学 | Resonant grounding system fault line selection method by combining complex wavelets with ANN (artificial neural network) |
| CN103884966A (en) * | 2014-04-15 | 2014-06-25 | 河海大学常州校区 | Power distribution network low-current single-phase earth fault positioning method based on neural network |
| CN105759167A (en) * | 2016-01-28 | 2016-07-13 | 江苏省电力公司南京供电公司 | Wavelet neural network-based distribution network single-phase short circuit line selection method |
| CN107340456A (en) * | 2017-05-25 | 2017-11-10 | 国家电网公司 | Power distribution network operating mode intelligent identification Method based on multiple features analysis |
-
2018
- 2018-01-30 CN CN201810088104.0A patent/CN108279364B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101924354A (en) * | 2010-04-19 | 2010-12-22 | 昆明理工大学 | Artificially neural network routing method for distribution network failure by using S-transforming energy sampling property |
| CN103163430A (en) * | 2013-03-29 | 2013-06-19 | 昆明理工大学 | Resonant grounding system fault line selection method by combining complex wavelets with ANN (artificial neural network) |
| CN103884966A (en) * | 2014-04-15 | 2014-06-25 | 河海大学常州校区 | Power distribution network low-current single-phase earth fault positioning method based on neural network |
| CN105759167A (en) * | 2016-01-28 | 2016-07-13 | 江苏省电力公司南京供电公司 | Wavelet neural network-based distribution network single-phase short circuit line selection method |
| CN107340456A (en) * | 2017-05-25 | 2017-11-10 | 国家电网公司 | Power distribution network operating mode intelligent identification Method based on multiple features analysis |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113945862A (en) * | 2021-10-18 | 2022-01-18 | 广东电网有限责任公司东莞供电局 | Method, device and equipment for identifying high-resistance grounding fault of power distribution network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108279364A (en) | 2018-07-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108279364B (en) | Power distribution network single-phase earth fault line selection method based on convolutional neural network | |
| Pourahmadi-Nakhli et al. | Path characteristic frequency-based fault locating in radial distribution systems using wavelets and neural networks | |
| CN103728535B (en) | A kind of extra-high-voltage direct-current transmission line fault location based on wavelet transformation transient state energy spectrum | |
| CN112041693A (en) | Power distribution network fault positioning system based on mixed wave recording | |
| CN109142851B (en) | Novel power distribution network internal overvoltage identification method | |
| CN114636900B (en) | Power distribution network multiple fault diagnosis method | |
| CN114441900A (en) | Power distribution network single-phase earth fault line selection method and system | |
| CN110596530A (en) | Low-current ground fault line selection method | |
| CN112240965A (en) | Grounding line selection device and method based on deep learning algorithm | |
| CN110247420B (en) | Intelligent fault identification method for HVDC transmission line | |
| CN112180217A (en) | Power distribution network single-phase earth fault section positioning method and system | |
| CN113821978B (en) | Traveling wave detection method and system based on improved step length LMS self-adaptive algorithm | |
| CN110726898A (en) | Power distribution network fault type identification method | |
| CN113985194A (en) | Power distribution network fault positioning method based on stack self-encoder | |
| CN114239703A (en) | Active power distribution system fault diagnosis method, system, equipment and storage medium | |
| Coban et al. | Artificial neural network based fault location on 230 kv transmission line using voltage and current signals | |
| CN115796033A (en) | Power transmission system fault diagnosis method and device, computer equipment and storage medium | |
| CN109782126B (en) | Power distribution network early fault detection method based on humanoid concept learning | |
| CN113158781B (en) | Lightning trip type identification method | |
| Yuan et al. | Faulty feeder detection for single phase-to-ground faults in distribution networks based on waveform encoding and waveform segmentation | |
| Wang et al. | Transmission line fault diagnosis based on wavelet packet analysis and convolutional neural network | |
| Yang et al. | Fault Classification in Distribution Systems using Deep Learning with Data Preprocessing Methods based on Fast-DTW and Short-Time Fourier Transform | |
| CN113610119A (en) | Method for identifying power transmission line developmental fault based on convolutional neural network | |
| CN112946425A (en) | Fault positioning method for mining travelling wave time-frequency domain characteristics by utilizing deep learning | |
| Kou et al. | Transmission line fault identification based on BP neural network |
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 |