CN116008731A - Power distribution network high-resistance fault identification method and device and electronic equipment - Google Patents
Power distribution network high-resistance fault identification method and device and electronic equipment Download PDFInfo
- Publication number
- CN116008731A CN116008731A CN202310119775.XA CN202310119775A CN116008731A CN 116008731 A CN116008731 A CN 116008731A CN 202310119775 A CN202310119775 A CN 202310119775A CN 116008731 A CN116008731 A CN 116008731A
- Authority
- CN
- China
- Prior art keywords
- fault
- feeder
- distribution network
- resistance
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 66
- 238000012549 training Methods 0.000 claims description 58
- 230000001052 transient effect Effects 0.000 claims description 32
- 238000004088 simulation Methods 0.000 claims description 25
- 238000010276 construction Methods 0.000 claims description 19
- 230000008569 process Effects 0.000 claims description 18
- 238000012545 processing Methods 0.000 claims description 18
- 238000010606 normalization Methods 0.000 claims description 12
- 238000004590 computer program Methods 0.000 claims description 10
- 239000004576 sand Substances 0.000 claims description 7
- 238000013526 transfer learning Methods 0.000 claims description 5
- 230000005540 biological transmission Effects 0.000 claims description 4
- 238000013135 deep learning Methods 0.000 claims description 4
- 238000013508 migration Methods 0.000 claims description 4
- 230000005012 migration Effects 0.000 claims description 4
- 238000010586 diagram Methods 0.000 description 11
- 238000012360 testing method Methods 0.000 description 10
- 230000006870 function Effects 0.000 description 6
- 238000004364 calculation method Methods 0.000 description 5
- 239000003990 capacitor Substances 0.000 description 5
- 238000003062 neural network model Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 239000003086 colorant Substances 0.000 description 3
- 230000005284 excitation Effects 0.000 description 3
- 230000001537 neural effect Effects 0.000 description 3
- 238000013473 artificial intelligence Methods 0.000 description 2
- 238000013145 classification model Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000005855 radiation Effects 0.000 description 2
- 239000004575 stone Substances 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008033 biological extinction Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000008094 contradictory effect Effects 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000009977 dual effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000010187 selection method Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
Images
Classifications
-
- 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
- Y04S10/52—Outage or fault management, e.g. fault detection or location
Landscapes
- Testing Of Short-Circuits, Discontinuities, Leakage, Or Incorrect Line Connections (AREA)
Abstract
The invention relates to a distribution network high-resistance fault identification method based on volt-ampere characteristic curve image identification, which comprises the following steps: acquiring feeder data of a feeder of a power distribution network; respectively constructing corresponding volt-ampere characteristic curves based on the acquired feeder line data; normalizing M voltammetric characteristic curves constructed based on feeder data of a single feeder into single feeder images, and arranging the M normalized single feeder images into a two-dimensional image serving as a fault identification image; and identifying the fault identification image according to a pre-constructed high-resistance fault identification model, and outputting a high-resistance fault identification result. According to the invention, the data of the head ends of all feeder lines of the power distribution network are utilized to construct the volt-ampere characteristic curve graph as the identification characteristic quantity of the high-resistance faults of the power distribution network, so that the problem of difficulty in characteristic quantity selection is effectively solved, meanwhile, the generalization capability of an identification model is effectively improved by combining the unbalanced sample set of the high-resistance fault causes, and the accuracy of identifying the high-resistance faults of the power distribution network is greatly improved.
Description
Technical Field
The invention relates to the technical field of power transmission line fault identification, in particular to a power distribution network high-resistance fault identification method and device based on volt-ampere characteristic curve image identification and electronic equipment.
Background
The power distribution network has wide and large range, complex running environment and easy damage caused by natural disasters and external force factors to cause faults. Meanwhile, the economic and social development has higher and higher requirements on safe and reliable operation of the power distribution network. Therefore, various faults in the power distribution network are detected rapidly and accurately, the fault types and reasons are identified, and the method has important significance for power distribution network fault recovery and operation and maintenance first-aid repair decisions.
The existing high-resistance fault identification research of the power distribution network is mainly focused on the detection of high-resistance faults, the high-resistance faults are detected under the influence of interference items and normal conditions, and the characteristics of the high-resistance faults are extracted by processing wave recording data or simulation data by combining a threshold method or an artificial intelligence method. The on-line monitoring method mainly comprises the following steps: (1) processing the feeder wave recording signal by using a signal processing method, and setting a threshold value as a basis for detecting the high-resistance fault; (2) and extracting the characteristics of the preprocessed feeder wave recording signals by using an artificial intelligence method, and identifying high-resistance faults. The simulation sample mainly considers interference items, normal conditions and general high-resistance faults.
The method is mainly used for identifying the high-resistance faults of a single fault cause, and has the following defects: (1) judging subjective influence exists in selection of the basis of the high-resistance faults, and objective rationality is lacking in evaluation indexes for feature quantity selection; (2) the threshold method is easily affected by noise, so that the effect of identifying the high-resistance faults is reduced; (3) the actual recording data is limited, the accuracy of the label is not ideal, and the effect of a fault classifier trained by using the data is not ideal; (4) the simulation acquisition of the fault sample does not consider various fault reasons of the high-resistance fault and the problem of unbalanced proportion among the fault reasons, and the trained model does not have good adaptability to actual working conditions. The results of training and testing the classification model are doubtful.
Therefore, it is needed to invent a power distribution network high-resistance fault identification method which combines a feature selection method capable of reflecting fault lines, phases and reasons and considers that the proportion of high-resistance fault reasons is unbalanced in actual conditions.
Disclosure of Invention
In view of the above, the invention provides a method, a device and an electronic device for identifying high-resistance faults of a power distribution network based on the image identification of a volt-ampere characteristic curve, which construct a volt-ampere characteristic curve graph by utilizing data of the head ends of all feeder lines of a power distribution network as the identification characteristic quantity of the high-resistance faults of the power distribution network, effectively solve the problem of difficult characteristic quantity selection, and simultaneously effectively improve the generalization capability of an identification model by combining a high-resistance fault cause type unbalanced sample set.
The invention discloses a power distribution network high-resistance fault identification method based on volt-ampere characteristic curve image identification, which comprises the following steps:
acquiring feeder data of a feeder of a power distribution network;
respectively constructing corresponding volt-ampere characteristic curves based on the acquired feeder line data;
normalizing M pieces of volt-ampere characteristic curves constructed based on feeder data of a single feeder into single feeder images, wherein M is more than or equal to 2, and arranging the M pieces of normalized single feeder images into a two-dimensional image serving as a fault identification image, wherein M is more than or equal to 1; specifically, a plurality of normalized single feeder images are transversely or longitudinally arranged to form a two-dimensional image;
And identifying the fault identification image according to a pre-constructed high-resistance fault identification model, and outputting a high-resistance fault identification result.
In the implementation process, the feeder data of the current power distribution network feeder is acquired first, then the acquired current feeder data is converted into a volt-ampere characteristic curve graph, and the volt-ampere characteristic curve graph acquired based on the current feeder data of the same feeder is processed into a single (one) feeder image through normalization, namely, the volt-ampere characteristic curve of the single feeder is generated and then displayed in an overlapping manner in one image. And combining the single feeder images of the multiple feeders into a new fault identification chart according to the feeder sequence in columns. Then, the fault identification image formed by the single feeder line image arrangement of the acquired single feeder line or multiple feeder lines (more than one feeder line) is identified through a pre-constructed high-resistance fault identification model (which can also be called a power distribution network fault identification classifier or is a part of the power distribution network fault identification classifier), and a high-resistance fault identification result is output after identification.
In a preferred embodiment, the fault identification result is output in the form of a fault tag field comprising a faulty line, a faulty phase and/or a high resistance fault cause.
In a preferred embodiment, the voltammetric characteristic graphs obtained based on the feeder data of the same feeder are distinguished in different colours.
Further, the feeder data comprise transient wave recording data of three-phase voltage and/or current signals of the feeder and transient wave recording data of zero-sequence voltage and/or current signals of the feeder; the constructing a corresponding volt-ampere characteristic curve based on the acquired feeder line data comprises the following steps: the method comprises the steps of constructing transient state recording data of three-phase voltage and/or current signals of a feeder line into a three-phase volt-ampere characteristic curve, and constructing transient state recording data of zero-sequence voltage and/or current signals of the feeder line into a zero-sequence volt-ampere characteristic curve.
Further, the normalizing the M volt-ampere characteristic curves constructed based on the feeder data of the single feeder into a single feeder image includes: normalizing a three-phase volt-ampere characteristic curve and a zero-sequence volt-ampere characteristic curve which are acquired based on a single feeder line into a single feeder line image; and combining the single feeder images of the multiple feeders into a new fault identification image according to the feeder sequence in columns.
In a preferred embodiment, the acquired feeder data are transient wave recording data of three-phase voltage and current signals and transient wave recording data of zero-sequence voltage and current signals at the head end of the feeder, the acquired transient wave recording data are then converted into volt-ampere characteristic curves corresponding to the acquired transient wave recording data, namely, a three-phase volt-ampere characteristic curve and a zero-sequence volt-ampere characteristic curve, and the three-phase volt-ampere characteristic curve and the zero-sequence volt-ampere characteristic curve of the same feeder are normalized to form a single feeder image. And combining the single feeder images of the multiple feeders into a new fault identification image according to the feeder sequence in columns. And the pre-constructed high-resistance fault recognition model recognizes the fault recognition image formed after normalization processing and outputs a recognition result. In a preferred embodiment, the three phase (A, B, C phase) and zero sequence volt-ampere characteristic curves of the feeder are distinguished in different colors.
In the invention, three-phase voltage and current and zero-sequence voltage and current at the head end of each feeder line of the power distribution network are used as inputs to construct and form a volt-ampere characteristic curve graph which is used as the identification characteristic quantity of the high-resistance fault of the power distribution network, thereby effectively solving the problem of difficult characteristic quantity extraction.
Further, before acquiring the feeder data of the power distribution network, the method further includes:
constructing an original recognition model;
acquiring a feeder line fault sample set;
inputting the obtained feeder line fault sample set into an original recognition model for transfer learning training until convergence, and obtaining a high-resistance fault recognition model.
Further, the original recognition model adopts a deep neural network model, including but not limited to CNN, alexNet, VGG16 and other neural networks. In a preferred embodiment, the original recognition model of the invention adopts an AlexNet network architecture, an input layer is constructed as a voltammetric characteristic curve image construction layer, an output layer is constructed as a fault recognition result output layer, and the rest layers are reserved. That is, in a preferred embodiment of the present invention, the voltammetric characteristic image constructing layer is part of the deep learning network model of the present invention (i.e., input layer) which replaces the input layer of the conventional AlexNet for respectively constructing the respective voltammetric characteristic based on the acquired feeder data and normalizing the respective voltammetric characteristic constructed based on the feeder data of the single feeder as the fault identification image. In further embodiments, the voltammetric curve image build layer can also be set independently of the employed network neural model, which is mainly dependent on the adaptation of the employed network neural model to the voltammetric curve image build layer.
In the implementation process, before real-time high-resistance fault identification is carried out, an original identification model is built, a feeder line fault sample set is obtained, and then the obtained feeder line fault sample set is input into the original identification model (a pre-trained AlexNet network model) for transfer learning training until convergence, so that a final power distribution network fault identification classifier is obtained. Specifically, after a feeder line fault sample set and an original recognition model to be trained are obtained, the fault sample set is randomly extracted to be used as a training sample set, the rest is used as a test sample set, after the normalization of training fault sample data is carried out, the training fault sample data is used as target domain data and is input into a pre-trained neural network model for training, the maximum training times are set, whether convergence is achieved or not is judged, training of the model is completed after the maximum training times are achieved on the premise of convergence, and if the training is not converged, the maximum training times are increased until convergence. And finally, inputting the accuracy of the test sample set calculation model to obtain a final high-resistance fault identification model.
In the implementation process, the deep learning network is trained by adopting the transfer learning method, so that a great amount of training time is saved, and the training speed is improved.
Further, obtaining a feeder fault sample set includes:
obtaining fault sample data used as a sample;
constructing a corresponding volt-ampere characteristic curve based on the acquired fault sample data, and carrying out normalization processing;
normalizing a corresponding volt-ampere characteristic curve constructed based on fault sample data of a single feeder line into a single feeder line fault image sample;
forming m normalized single feeder line fault image samples into a two-dimensional image serving as a feeder line fault sample according to the arrangement of columns, wherein m is more than or equal to 2;
a plurality of fault samples are acquired to form a feeder fault sample set.
Further, the fault sample set is an unbalanced-like sample set generated based on different high-resistance fault causes, wherein the high-resistance fault causes comprise a touch tree-like fault, an animal-like fault and/or a sand-like fault. In a preferred embodiment, in the present invention, the high resistance fault causes include three of a trigger tree fault, an animal fault, and a sand fault. It will be appreciated by those skilled in the relevant arts that the high resistance fault causes can include, but are not limited to, three types of faults including a trigger tree fault, an animal fault, and a sand fault, and that other types of fault causes are equally applicable to the disclosed methods.
Further, the fault sample data is obtained by constructively simulating the sample according to actual distribution network parameters, including:
according to the high-resistance fault recording data of the actual power distribution network, the probability distribution among different reasons, phases, types, distances and phase angles of the high-resistance faults in the power distribution network is counted;
constructing simulation samples corresponding to different high-resistance fault reasons according to the high-resistance fault recording data of the actual power distribution network, wherein the simulation samples corresponding to the different high-resistance fault reasons respectively meet probability distribution among different reasons, fault types, distances and phase angles of the high-resistance fault;
fault sample data corresponding to a cause of a fault is generated based on the simulation sample.
Further, in the step of constructing simulation samples corresponding to different high-resistance fault reasons according to the high-resistance fault recording data of the actual power distribution network, when the actual high-resistance fault recording data meets the quantity condition, constructing the simulation samples by utilizing the actual wave recording waveforms; when the actual high-resistance fault recording data does not meet the quantity condition, referring to the actual recording waveform, simulating the actual recording waveform in a modeling mode to construct a simulation sample. The number condition is set as follows: the number of recording data is 10000.
In the implementation process, modeling is performed by referring to actual recording data aiming at different fault reasons, so that simulation of high-resistance faults of different fault reasons is realized, the influence of the fault reasons is increased in construction of a fault sample set, and identification of the high-resistance fault reasons of the power distribution network is realized. In further embodiments, interference influence terms, including excitation capacity, capacitance switching, load variation, and the like, are also considered in the construction of the fault sample set, which further increases the identification of the cause of the high-resistance fault of the power distribution network.
In addition, in the implementation process, by means of the statistical probability distribution of the fault types, the fault phases and the fault reasons in the actual fault records, the fault sample set conforming to the actual fault records can be generated in batches, the training of the fault identification classifier is facilitated, and the generalization capability and the application effect of the fault identification classifier to the actual conditions are improved.
Further, the fault sample data includes transient wave recording data of three-phase voltage and/or current signals of the feeder samples, and transient wave recording data of zero-sequence voltage and/or current signals of the feeder samples. In a preferred embodiment, the fault sample data is transient log data of three-phase voltage and current signals of the feeder samples, and transient log data of zero-sequence voltage and current signals of the feeder samples.
Further, the step of respectively constructing corresponding volt-ampere characteristic curves based on the acquired fault sample data and performing normalization processing includes: the transient state recording data of the three-phase voltage and/or current signals of the feeder line sample are constructed into a three-phase volt-ampere characteristic curve, and the transient state recording data of the zero-sequence voltage and/or current signals of the feeder line sample are constructed into a zero-sequence volt-ampere characteristic curve.
Further, normalizing the corresponding volt-ampere characteristic curve constructed based on the fault sample data of the single feeder line sample into a single feeder line fault image sample, comprising: and normalizing the three-phase volt-ampere characteristic curve and the zero-sequence volt-ampere characteristic curve which are obtained based on the single feeder line sample into a single feeder line fault image sample.
In the implementation process, three-phase voltage, current, zero-sequence voltage and current at the head end of a feeder line of the power distribution network are converted into a three-phase volt-ampere characteristic curve and a zero-sequence volt-ampere characteristic curve, the obtained volt-ampere characteristic curve is normalized into a single feeder line fault image sample, and then a plurality of feeder line fault image samples are integrated into a two-dimensional image according to the arrangement of columns to serve as input quantity of an original recognition model, so that migration learning training is conducted on the original recognition model until convergence is achieved. By taking the two-dimensional image as the input quantity of the recognition model, the physical meaning of the input quantity can be reflected, and the extraction of fault characteristics is facilitated.
Further, the feeder data and the fault sample data are data obtained from the feeder head end. In other embodiments, the feeder data and fault sample data can be data obtained from the end of the feeder, or can be data obtained from the head end and end of the feeder.
Further, in the generation process of the feeder fault sample, the per unit value of the data acquired by the feeder head end is calculated according to the voltage level, the line parameter and the transmission capacity of the power distribution network, and the following principle is followed:
a. the three-phase voltage signals at the head end of the feeder line of the power distribution network adopt the same reference value;
b. the three-phase current signals at the head end of the feeder line of the power distribution network adopt the same reference value;
c. the zero sequence voltage signals at the head end of the feeder line of the power distribution network adopt the same reference value;
d. the zero sequence current signals at the head end of the feeder line of the power distribution network adopt the same reference value;
e. the abscissa range is defined as [ -1,1].
The second aspect of the invention also discloses a distribution network high-resistance fault identification device based on the voltammetric characteristic curve image identification, which comprises the following components:
the first acquisition unit is used for acquiring feeder line data of the power distribution network;
the volt-ampere characteristic curve construction unit is used for respectively constructing corresponding volt-ampere characteristic curves based on the acquired feeder line data;
The image processing unit normalizes M pieces of volt-ampere characteristic curves constructed based on feeder data of a single feeder into single feeder images, wherein M is more than or equal to 2, and the M normalized single feeder images are arranged to form a two-dimensional image serving as a fault identification image, wherein M is more than or equal to 1;
the fault identification and classification unit is used for identifying the obtained fault identification image according to a pre-constructed high-resistance fault identification model and outputting a high-resistance fault identification result.
In the implementation process, the first acquisition unit acquires the feeder data of the current power distribution network, the volt-ampere characteristic curve construction unit constructs a volt-ampere characteristic curve corresponding to the data based on the acquired feeder data, and the image processing unit converts the volt-ampere characteristic curve constructed by the volt-ampere characteristic curve construction unit into a two-dimensional image serving as the identification characteristic quantity of the fault identification classification unit, so that the problem of difficulty in characteristic quantity selection is effectively solved. In this embodiment of the invention, the voltammetric curve construction unit and the image processing unit together form a voltammetric curve construction layer, which can be part of the recognition model.
Further, the power distribution network high-resistance fault identification device further comprises:
The model construction unit is used for constructing an original identification model before acquiring the feeder line fault sample set data;
the second acquisition unit is used for acquiring a feeder line fault sample set;
the training unit is used for performing migration learning training on the original recognition model through the obtained feeder line fault sample set until convergence to obtain a high-resistance fault recognition model.
In the implementation process, the training unit divides the feeder line fault sample set into a training sample set and a test sample set, after the training fault sample data is normalized, the training fault sample data is used as target domain data to be input into a pre-trained neural network model for training, the maximum training times are set, whether the training is converged or not is judged, the training of the model is completed after the maximum training times are reached on the premise of convergence, and the maximum training times are increased until the training is converged if the training is not converged. And finally, inputting the accuracy of the test sample set calculation model to obtain a final high-resistance fault identification model.
The third aspect of the invention also discloses an electronic device, which comprises a memory and a processor, wherein the memory is used for storing a computer program, and the processor runs the computer program to enable the electronic device to execute the power distribution network high-resistance fault identification method based on the voltammetric characteristic curve image identification disclosed in the first aspect of the invention.
The fourth aspect of the present invention also discloses a computer readable storage medium, which stores computer program instructions, and when the computer program instructions are read and executed by a processor, the method for identifying high-resistance faults of a power distribution network based on the voltage-current characteristic curve image identification in the first aspect of the present invention is executed.
Compared with the prior art, the distribution network high-resistance fault identification method based on the voltammetric characteristic curve image identification has the following beneficial effects:
according to the invention, the acquired feeder line data are converted into the volt-ampere characteristic curve, and the volt-ampere characteristic curve is converted into the two-dimensional image to be used as a fault recognition graph, so that the recognition characteristic quantity of the high-resistance fault of the power distribution network is formed, the problem of difficulty in characteristic quantity selection is effectively solved, and the problem of lack of objective rationality in judging the selection of the characteristic quantity of the high-resistance fault in the prior art is overcome; and a two-dimensional image based on the volt-ampere characteristic curve is used as the identification characteristic quantity, so that the problem that the high-resistance fault identification effect is reduced due to the fact that the threshold method is easily affected by noise in the prior art is avoided.
The invention discloses a power distribution network high-resistance fault identification method, a device and electronic equipment based on voltammetric characteristic curve image identification in detail with reference numerals in combination with the embodiment shown in the drawings.
Drawings
Fig. 1 shows a flow chart of steps of a high-resistance fault identification method of a power distribution network based on voltammetric characteristic curve image identification in the invention.
Fig. 2 shows a flowchart of the steps for pre-constructing a high-resistance fault recognition model in the present invention.
Fig. 3 shows a schematic architecture diagram of a high-resistance fault recognition model in the present invention.
Fig. 4 shows an example of a two-dimensional image formed by normalizing and arranging a three-phase volt-ampere characteristic curve and a zero-sequence volt-ampere characteristic curve of a feeder line of a power distribution network.
Fig. 5 is a schematic diagram of a simulation model of a single-power distribution network in an embodiment of the invention.
FIG. 6 is a schematic diagram of a dual loop tower model in an embodiment of the present invention.
FIG. 7 is a schematic diagram of a tree-type high-resistance fault model in an embodiment of the invention.
FIG. 8 is a schematic diagram of an animal high resistance fault model in an embodiment of the invention.
FIG. 9 is a schematic diagram of a sand-type high-resistance fault model in an embodiment of the invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
It should be noted that, in the embodiments of the present invention, all directional indicators (such as up, down, left, right, front, and rear … …) are merely used to explain the relative positional relationship, movement conditions, and the like between the components in a specific posture (as shown in the drawings), and if the specific posture is changed, the directional indicators are correspondingly changed.
Furthermore, the description of "first," "second," etc. in this disclosure is for descriptive purposes only and is not to be construed as indicating or implying a relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include at least one such feature. In addition, the technical solutions of the embodiments may be combined with each other, but it is necessary to base that the technical solutions can be realized by those skilled in the art, and when the technical solutions are contradictory or cannot be realized, the combination of the technical solutions should be regarded as not exist and not within the protection scope of the present invention.
The invention discloses a high-resistance fault identification method of a power distribution network based on voltammetric characteristic curve image identification, shown in fig. 1, comprising a high-resistance fault identification model pre-constructed based on a deep learning network model, wherein the method specifically comprises the following steps:
Acquiring feeder data of a feeder of a power distribution network;
respectively constructing corresponding volt-ampere characteristic curves based on the acquired feeder line data;
normalizing M pieces of volt-ampere characteristic curves constructed based on feeder data of a single feeder into single feeder images, wherein M is more than or equal to 2, and arranging the M pieces of normalized single feeder images into a two-dimensional image serving as a fault identification image, wherein M is more than or equal to 1;
and identifying the fault identification image according to a pre-constructed high-resistance fault identification model, and outputting a high-resistance fault identification result.
In one embodiment, the feeder data includes transient log data of three-phase voltage and/or current signals of the feeder and transient log data of zero-sequence voltage and/or current signals of the feeder. The method comprises the steps of respectively constructing corresponding volt-ampere characteristic curves based on acquired feeder data, normalizing the corresponding volt-ampere characteristic curves constructed based on the feeder data of single feeder into single feeder images, and comprises the following steps: constructing transient recording data of three-phase voltage and/or current signals of a feeder line into a three-phase volt-ampere characteristic curve; constructing transient state recording data of zero sequence voltage and/or current signals of a feeder line into a zero sequence volt-ampere characteristic curve; and normalizing the three-phase volt-ampere characteristic curve and the zero-sequence volt-ampere characteristic curve which are acquired based on the single feeder line into a single feeder line image.
As shown in fig. 2, in the above embodiment, the process of constructing the high-resistance fault identification model in the present invention includes the following steps:
constructing an original recognition model;
acquiring a feeder line fault sample set;
inputting the obtained feeder line fault sample set into the original recognition model for transfer learning training until convergence, and obtaining the high-resistance fault recognition model.
In one embodiment, the step of obtaining a feeder fault sample set includes:
obtaining fault sample data used as a sample;
constructing a corresponding volt-ampere characteristic curve based on the acquired fault sample data, and carrying out normalization processing;
normalizing a corresponding volt-ampere characteristic curve constructed based on fault sample data of a single feeder line into a single feeder line fault image sample;
forming m normalized single feeder line fault image samples into a two-dimensional image serving as a feeder line fault sample according to the arrangement of columns, wherein m is more than or equal to 1;
a plurality of the fault samples are acquired to form a feeder fault sample set.
In the above embodiment, the fault sample set is a class imbalance sample set generated based on different high-resistance fault causes including at least two of a trigger tree type fault, an animal type fault, and a sand type fault.
In the above embodiment, the fault sample data includes transient log data of three-phase voltage and/or current signals of the feeder samples, and transient log data of zero-sequence voltage and/or current signals of the feeder samples.
In the foregoing embodiment, the steps of respectively constructing the corresponding volt-ampere characteristic curves based on the obtained fault sample data and performing normalization processing include: the transient state recording data of the three-phase voltage and/or current signals of the feeder line sample are constructed into a three-phase volt-ampere characteristic curve, and the transient state recording data of the zero-sequence voltage and/or current signals of the feeder line sample are constructed into a zero-sequence volt-ampere characteristic curve.
In the above embodiment, the step of normalizing the corresponding voltammetric characteristic curve constructed based on the fault sample data of the single feeder sample into the single feeder fault image sample includes: and normalizing the three-phase volt-ampere characteristic curve and the zero-sequence volt-ampere characteristic curve which are obtained based on the single feeder line sample into a single feeder line fault image sample.
In the above embodiment, in the generation process of the feeder fault sample, the per unit value of the data acquired by the feeder head end is calculated according to the voltage level, the line parameter and the transmission capacity of the power distribution network, which follows the following principle:
a. The three-phase voltage signals at the head end of the feeder line of the power distribution network adopt the same reference value;
b. the three-phase current signals at the head end of the feeder line of the power distribution network adopt the same reference value;
c. the zero sequence voltage signals at the head end of the feeder line of the power distribution network adopt the same reference value;
d. the zero sequence current signals at the head end of the feeder line of the power distribution network adopt the same reference value;
e. the abscissa range is defined as [ -1,1].
Fig. 3 shows a schematic architecture diagram of a high-resistance fault recognition model in the present invention. Fig. 4 shows an example of a two-dimensional image formed by normalizing and arranging a three-phase volt-ampere characteristic curve and a zero-sequence volt-ampere characteristic curve of a feeder line of a power distribution network. As shown in fig. 3 and 4, in the voltammetric characteristic curve construction layer, a plurality of voltammetric characteristic curves of a single feeder line are displayed in an overlapping manner in the same coordinate system. In a preferred embodiment, the multiple voltammetric characteristics of a single feed line are distinguished by different colours. The voltammetric characteristic curve images of a plurality of single feeder lines are subgraphs, and the feeder lines are combined into a new voltammetric characteristic curve image according to the sequence of the feeder lines. All voltammetric characteristic curves are plotted in the form of per unit value, and the occupied area of each subgraph is the same.
In the above embodiment, the waveform of all the transient recording data takes three periods.
In the above embodiment, the original recognition model adopts an AlexNet network architecture, the input layer of which is configured as the voltammetric characteristic curve image building layer, and the output layer is configured as the fault recognition result output layer. The original recognition model is trained and tested by a fault sample set to finally form the high-resistance fault recognition model. Specifically, in the embodiment of the invention, a pre-trained AlexNet network model is used, an original image is input and changed into a 227×227×3 image, a pixel layer is formed through convolution kernel convolution calculation, then an activation function ReLU, a local response normalization Layer (LRN) and a maximum pooling layer enter the next convolution layer, the operation is repeated for 5 times, a 6 th full-connection network layer is reached, the operation is output to the next layer through a ReLU function and a dropout operation, and the operation is repeated for 2 times, and then the operation reaches an 8 th layer. The parameters and the weights of the pre-training AlexNet network are reserved, and the original output layer is replaced by a new output layer; the full-connection layer of the classified output layer integrates the feature matrix processed by the middle layer to form feature quantity, and the quantity with highest possibility is determined by the Softmax layer to be output as a result.
In the above embodiment, the high-resistance fault identification model (i.e., the deep neural learning network model) of the present invention can output N types of structures, where the method for determining N according to the combination of fault line-fault phase-high-resistance fault cause is as follows:
Wherein: for a general single-power radiation type power distribution network system, N HIF Represents the number of high-resistance fault categories, f represents the number of feeder lines of the system, m represents the number of phases of the system, namely a, b and c 3 phases, I HIF The number of high-resistance fault cause categories is represented; h CCF Representing the number of non-single-phase earth fault types; n (N) D Representing the number of interference categories, I D Indicating the number of interference cause categories.
For the simulation model of the single-power distribution network as shown in fig. 5, the high-resistance fault of the invention needs to identify the touch tree, the sand and the animal 3, then I HIF =3; the system has 4 feeder lines and 3 feeder lines, f=4 and m=3; considering the line crossing faults of A, C phases of double-circuit lines on the same pole, the fault reasons comprise tree contact and animal electric shock, so H CCF =4; to sum up, the number of the high-resistance fault categories to be identified is 40; for interference items, three interference items of capacitance switching, exciting inrush current and load switching are considered, I D =3, f=4; the number of the types of the interference items to be identified is 12, and 52 types are identified altogether.
In the above embodiment, the fault sample data is obtained by constructively simulating the sample according to the actual distribution network parameters, including: according to the high-resistance fault recording data of the actual power distribution network, the probability distribution among high-resistance faults of different reasons in the power distribution network is counted; constructing simulation samples corresponding to different high-resistance fault reasons according to the high-resistance fault recording data of the actual power distribution network, wherein the simulation samples corresponding to the different high-resistance fault reasons meet probability distribution; fault sample data corresponding to a cause of a fault is generated based on the simulation sample.
In the step of constructing simulation samples corresponding to different high-resistance fault reasons according to the high-resistance fault recording data of the actual power distribution network, when the actual high-resistance fault recording data meets the quantity conditions, constructing the simulation samples by utilizing the actual wave recording waveforms; when the actual high-resistance fault recording data does not meet the quantity condition, referring to the actual recording waveform, simulating the actual recording waveform in a modeling mode to construct a simulation sample.
Fig. 5 shows an example of a simulation model of a power distribution network that constructs a set of fault samples in a simulated manner. In fig. 5, referring to a single power source radiation system of an actual power distribution network, T is a main transformer, 4 feeder lines are all overhead lines, L0 represents a same-pole four-circuit line, L3 represents a same-pole double-circuit line, and L1, L2, L4 and L5 are single-circuit lines. The line model was 30/7ACSR, the outer radius of the wire was 0.933cm, and the resistance was 0.135 Ω/km. F1-F6 and F10-F15 are selected single-phase high-resistance grounding fault points, and the interval is 0.225km; F7-F9 are set double-circuit line overline fault points, and the distance is 0.15km. The current transformer and the voltage transformer are arranged at the head end of each feeder line.
FIG. 6 shows an example of a pole and tower model in a simulation model for constructing a set of fault samples in a simulated manner.
When the high-resistance fault cause type unbalanced sample is constructed in the embodiment, three fault causes are respectively a trigger tree type fault, an animal type fault and a sand stone type fault.
Wherein, the modeling of the contact tree fault is based on the breakdown characteristic of the solid medium, as shown in fig. 5, the relation between the current and the voltage in the medium is fitted to a piecewise curve to reflect the resistor R in the contact tree high-resistance fault transition impedance tf The nonlinear variation characteristic of the system enables the fault impedance angle to randomly vary within the range of-10 degrees to 10 degrees at the frequency of 10kHz.
The parameters of the conductance portion of the impedance are determined by the following formula:
wherein: i is fault current; u is the fault voltage; g 1 ，g 2 Conductance values for different phases; k (k) 1 、k 2 、k 3 K 4 Is a coefficient to be determined.
The model parameters were set as: u (u) 1 Set to 2kV, u 2 Set to 8kV, R ft1 Set to 8000 Ω, R ft2 Set to 1000 Ω, k 1 、k 2 、k 3 K 4 According to the value of g tf The conductance value is continuously calculated.
Wherein, modeling of animal faults divides the animal electric shock impedance into internal and external impedance, as shown in figure 8The resistance takes a fixed value of 500 omega, the external impedance consists of a resistor and a capacitor which are connected in series, the fluctuation range of the resistor is 1000 omega-100 kΩ, and the capacitor Cs takes a fixed value of 20nF/cm 2 The frequency of variation was 10kHz.
In order to simulate the transient arcing process of animal high-resistance faults, animal electric shock impedance is connected in series with tree-contact fault impedance resistor R ft Obtaining the animal high-resistance fault impedance.
The structure of the sandstone type high-resistance ground fault model is shown in figure 7, the model comprises two nonlinear variable resistors used for representing nonlinearity of fault current, and two nonlinear direct current power supplies connected by diodes are connected in anti-parallel to simulate asymmetric distortion property of the fault current and arc extinction phenomenon near a current zero point, and the change frequency is 10kHz.
In the embodiment, the interference term considers three types of excitation surge current, capacitor switching and load switching;
excitation surge current interference samples are generated through opening and closing of a transformer disconnecting link;
the capacitor switching interference sample is generated by switching a capacitor at a fault position;
load switching disturbance samples are generated by varying feeder end load parameters.
Thus, the present example generates a total of 5000 groups of high resistance fault samples, 1320 groups of interference samples; among the high-resistance fault samples, 3000 groups (60%) of the trigger tree type high-resistance fault samples, 2000 groups (40%) of the sand stone type high-resistance fault samples and 500 groups (10%) of the animal type high-resistance fault samples are adopted.
In the aspect of fault phase, non-line crossing fault phase is uniformly generated; in the line crossing fault aspect, in the L3 double circuit line setting as illustrated in fig. 5, referring to the tower model of the same-pole double circuit line as illustrated in fig. 6, a phase and a phase C line crossing faults are set, and fault sample data is generated by randomly changing model parameters, uniformly changing fault point positions and starting phase angle simulation of fault occurrence.
After obtaining a fault sample and a fault identification model to be trained, randomly extracting 70% of the fault sample as a training sample set, taking the rest as a test sample set, normalizing training sample data, inputting the normalized training sample data as target domain data to a pre-training neural network model for training, setting the maximum training times, judging whether convergence, completing training of the model after reaching the maximum training times on the premise of convergence, and increasing the maximum training times until convergence if not. And finally, inputting the accuracy of the test sample set calculation model to obtain a final classification model.
The test results are shown in tables 1 and 2, and the results show that the accuracy can reach 100% both in fault type identification and fault cause identification.
TABLE 1 accuracy of identification of high resistance failure types by the method of the present invention
Note that: single digit and single phase combinations represent faulty lines and phases, e.g., 1A represents a 1-line a-phase fault; double numbers and double phases indicate that the line and the phase of the single tower double circuit line are located, for example, 3A and 4A indicate that the double circuit line with 3 lines and 4 lines has a phase A cross line fault.
TABLE 2 accuracy of identification of cause of high resistance failures by the method of the present invention
The embodiment of the invention also provides a power distribution network high-resistance fault identification device based on the voltammetric characteristic curve image identification, which comprises the following steps:
The first acquisition unit is used for acquiring feeder line data of the power distribution network;
the volt-ampere characteristic curve construction unit is used for respectively constructing corresponding volt-ampere characteristic curves based on the acquired feeder line data;
the image processing unit normalizes M pieces of volt-ampere characteristic curves constructed based on feeder data of a single feeder into single feeder images, wherein M is more than or equal to 2, and the M normalized single feeder images are arranged to form a two-dimensional image serving as a fault identification image, wherein M is more than or equal to 1;
the fault identification and classification unit is used for identifying the obtained fault identification image according to a pre-constructed high-resistance fault identification model and outputting a high-resistance fault identification result.
In the implementation process, the first acquisition unit acquires the feeder data of the current power distribution network, the volt-ampere characteristic curve construction unit constructs a volt-ampere characteristic curve corresponding to the data based on the acquired feeder data, and the image processing unit converts the volt-ampere characteristic curve constructed by the volt-ampere characteristic curve construction unit into a two-dimensional image serving as the identification characteristic quantity of the fault identification classification unit, so that the problem of difficulty in characteristic quantity selection is effectively solved. In this embodiment of the invention, the voltammetric curve construction unit and the image processing unit together form a voltammetric curve construction layer, which can be part of the recognition model.
In an embodiment of the foregoing apparatus, the power distribution network high-resistance fault identification apparatus further includes:
the model construction unit is used for constructing an original identification model before acquiring the feeder line fault sample set data;
the second acquisition unit is used for acquiring a feeder line fault sample set;
the training unit is used for performing migration learning training on the original recognition model through the obtained feeder line fault sample set until convergence to obtain a high-resistance fault recognition model.
In the implementation process, the training unit divides the feeder line fault sample set into a training sample set and a test sample set, after the training fault sample data is normalized, the training fault sample data is used as target domain data to be input into a pre-trained neural network model for training, the maximum training times are set, whether the training is converged or not is judged, the training of the model is completed after the maximum training times are reached on the premise of convergence, and the maximum training times are increased until the training is converged if the training is not converged. And finally, inputting the accuracy of the test sample set calculation model to obtain a final high-resistance fault identification model.
The embodiment of the invention also discloses electronic equipment, which comprises a memory and a processor, wherein the memory is used for storing a computer program, and the processor runs the computer program to enable the electronic equipment to execute the power distribution network high-resistance fault identification method based on the voltammetric characteristic curve image identification disclosed in the first aspect of the invention.
The embodiment of the invention also discloses a computer readable storage medium which stores computer program instructions, and when the computer program instructions are read and run by a processor, the method for identifying the high-resistance faults of the power distribution network based on the voltage-current characteristic curve image identification in the first aspect of the embodiment of the invention is executed.
In the several embodiments provided in this application, it should be understood that the disclosed apparatus and method may be implemented in other manners as well. The apparatus embodiments described above are merely illustrative, for example, of the flowcharts and block diagrams in the figures that illustrate the architecture, functionality, and operation of possible implementations of methods according to embodiments of the present application. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
In addition, the functional modules in the embodiments of the present application may be integrated together to form a single part, or each module may exist alone, or two or more modules may be integrated to form a single part.
The functions, if implemented in the form of software functional modules and sold or used as a stand-alone product, may be stored in a computer-readable storage medium. Based on such understanding, the technical solution of the present application may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
The foregoing is merely exemplary embodiments of the present application and is not intended to limit the scope of the present application, and various modifications and variations may be suggested to one skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application. It should be noted that: like reference numerals and letters denote like items in the following figures, and thus once an item is defined in one figure, no further definition or explanation thereof is necessary in the following figures.
The foregoing is merely specific embodiments of the present application, but the scope of the present application is not limited thereto, and any person skilled in the art can easily think about changes or substitutions within the technical scope of the present application, and the changes and substitutions are intended to be covered by the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
It is noted that relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
Claims (10)
1. A power distribution network high-resistance fault identification method based on volt-ampere characteristic curve image identification is characterized by comprising the following steps:
acquiring feeder data of a feeder of a power distribution network;
constructing a corresponding volt-ampere characteristic curve based on the acquired feeder line data;
normalizing M pieces of volt-ampere characteristic curves constructed based on feeder data of a single feeder into single feeder images, wherein M is more than or equal to 2, and arranging the M pieces of normalized single feeder images into a two-dimensional image serving as a fault identification image, wherein M is more than or equal to 1;
and identifying the fault identification image according to a pre-constructed high-resistance fault identification model, and outputting a high-resistance fault identification result.
2. The method for identifying the high-resistance fault of the power distribution network based on the voltammetric characteristic curve image recognition according to claim 1, wherein the feeder data comprises transient wave recording data of three-phase voltage and/or current signals of a feeder and transient wave recording data of zero-sequence voltage and/or current signals of the feeder;
the constructing a corresponding volt-ampere characteristic curve based on the acquired feeder line data comprises the following steps: constructing transient state recording data of three-phase voltage and/or current signals of the feeder line into a three-phase volt-ampere characteristic curve, and constructing transient state recording data of zero-sequence voltage and/or current signals of the feeder line into a zero-sequence volt-ampere characteristic curve;
The normalization processing of the M volt-ampere characteristic curves constructed based on the feeder data of the single feeder into the single feeder image comprises the following steps: and normalizing the three-phase volt-ampere characteristic curve and the zero-sequence volt-ampere characteristic curve which are acquired based on the single feeder line into a single feeder line image.
3. The method for identifying high-resistance faults of a power distribution network based on voltammetric characteristic image recognition according to claim 1 or 2, wherein before obtaining feeder data of the power distribution network, the method further comprises:
constructing an original recognition model;
acquiring a feeder line fault sample set;
inputting the obtained feeder line fault sample set into the original recognition model for transfer learning training until convergence to obtain the high-resistance fault recognition model;
wherein, the obtaining a feeder fault sample set includes:
obtaining fault sample data used as a sample;
constructing a corresponding volt-ampere characteristic curve based on the acquired fault sample data, and carrying out normalization processing;
normalizing a corresponding volt-ampere characteristic curve constructed based on fault sample data of a single feeder line into a single feeder line fault image sample;
forming m normalized single feeder line fault image samples into a two-dimensional image serving as a feeder line fault sample according to the arrangement of columns, wherein m is more than or equal to 1;
Obtaining a plurality of feeder fault samples to form a feeder fault sample set.
4. The method for identifying high-resistance faults of a power distribution network based on voltammetric characteristic curve image recognition according to claim 3, wherein the fault sample set is an unbalanced-like sample set generated based on different high-resistance fault causes, wherein the high-resistance fault causes comprise contact tree faults, animal faults and/or sand faults;
the fault sample data is obtained through a simulation sample constructed according to actual power distribution network parameters, and the fault sample data comprises the following components:
according to the high-resistance fault recording data of the actual power distribution network, the probability distribution among different reasons, phases, types, distances and phase angles of the high-resistance faults in the power distribution network is counted; constructing simulation samples corresponding to different high-resistance fault reasons according to the high-resistance fault recording data of the actual power distribution network, wherein the simulation samples corresponding to the different high-resistance fault reasons respectively meet probability distribution among different reasons, fault types, distances and phase angles of the high-resistance fault;
fault sample data corresponding to a cause of a fault is generated based on the simulation sample.
5. The method for identifying the high-resistance faults of the power distribution network based on the voltammetric characteristic curve image recognition according to claim 3, wherein the fault sample data comprises transient wave recording data of three-phase voltage and/or current signals of feeder samples and transient wave recording data of zero-sequence voltage and/or current signals of feeder samples;
The method comprises the steps of respectively constructing corresponding volt-ampere characteristic curves based on acquired fault sample data, and carrying out normalization processing, wherein the steps comprise: constructing transient state recording data of three-phase voltage and/or current signals of a feeder line sample into a three-phase volt-ampere characteristic curve, and constructing transient state recording data of zero-sequence voltage and/or current signals of the feeder line sample into a zero-sequence volt-ampere characteristic curve;
the normalization processing of the corresponding volt-ampere characteristic curve constructed based on the fault sample data of the single feeder line sample into the single feeder line fault image sample comprises the following steps: and normalizing the three-phase volt-ampere characteristic curve and the zero-sequence volt-ampere characteristic curve which are obtained based on the single feeder line sample into a single feeder line fault image sample.
6. The method for identifying the high-resistance faults of the power distribution network based on the voltammetric characteristic curve image identification according to claim 3, wherein the original identification model adopts a pre-trained deep learning network architecture, an input layer is constructed as a voltammetric characteristic curve image construction layer, and an output layer is constructed as a fault identification result output layer.
7. The method for identifying high-resistance faults of a power distribution network based on voltammetric characteristic curve image recognition according to claim 3, wherein the feeder line data and the fault sample data are data obtained from a head end of a feeder line;
In the generation process of the feeder fault sample, calculating the per unit value of the data acquired by the feeder head end according to the voltage level, the line parameter and the transmission capacity of the power distribution network, wherein the per unit value follows the following principle:
a. the three-phase voltage signals at the head end of the feeder line of the power distribution network adopt the same reference value;
b. the three-phase current signals at the head end of the feeder line of the power distribution network adopt the same reference value;
c. the zero sequence voltage signals at the head end of the feeder line of the power distribution network adopt the same reference value;
d. the zero sequence current signals at the head end of the feeder line of the power distribution network adopt the same reference value;
e. the abscissa range is defined as [ -1,1].
8. The utility model provides a distribution network high resistance trouble recognition device based on volt-ampere characteristic curve image recognition which characterized in that, distribution network high resistance trouble recognition device includes:
the first acquisition unit is used for acquiring feeder line data of the power distribution network;
the volt-ampere characteristic curve construction unit is used for respectively constructing corresponding volt-ampere characteristic curves based on the acquired feeder line data;
the image processing unit normalizes M pieces of volt-ampere characteristic curves constructed based on feeder data of a single feeder into single feeder images, wherein M is more than or equal to 2, and the M normalized single feeder images are arranged to form a two-dimensional image serving as a fault identification image, wherein M is more than or equal to 1;
The fault identification and classification unit is used for identifying the single feeder line image obtained by normalizing the single feeder line according to a pre-constructed high-resistance fault identification model and outputting a high-resistance fault identification result.
9. The distribution network high-resistance fault identification device based on the voltammetric characteristic image recognition according to claim 8, wherein the distribution network high-resistance fault identification device further comprises:
the model construction unit is used for constructing an original identification model before the feeder line fault sample set data are acquired;
the second acquisition unit is used for acquiring a feeder line fault sample set;
and the training unit is used for performing migration learning training on the original recognition model through the obtained feeder line fault sample set until convergence to obtain the high-resistance fault recognition model.
10. An electronic device comprising a memory for storing a computer program and a processor that runs the computer program to cause the electronic device to perform the method of identifying a high resistance fault of a power distribution network based on voltammetric characteristic image recognition according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310119775.XA CN116008731B (en) | 2023-02-15 | 2023-02-15 | Power distribution network high-resistance fault identification method and device and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310119775.XA CN116008731B (en) | 2023-02-15 | 2023-02-15 | Power distribution network high-resistance fault identification method and device and electronic equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116008731A true CN116008731A (en) | 2023-04-25 |
| CN116008731B CN116008731B (en) | 2023-08-25 |
Family
ID=86033654
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310119775.XA Active CN116008731B (en) | 2023-02-15 | 2023-02-15 | Power distribution network high-resistance fault identification method and device and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116008731B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116703284A (en) * | 2023-08-03 | 2023-09-05 | 八爪鱼人工智能科技(常熟)有限公司 | Fault identification method applied to refrigeration house management system and artificial intelligent server |
| CN117290756A (en) * | 2023-09-25 | 2023-12-26 | 重庆大学 | Power transmission line fault identification method and device based on federal learning and electronic equipment |
Citations (24)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2013024924A1 (en) * | 2011-08-16 | 2013-02-21 | 한국전력공사 | Apparatus and method for detecting a fault section of a power distribution system |
| US20150226780A1 (en) * | 2014-02-07 | 2015-08-13 | Mitsubishi Electric Research Laboratories, Inc. | Locating Multi-Phase Faults in Ungrounded Power Distribution Systems |
| US20160041216A1 (en) * | 2013-03-29 | 2016-02-11 | Beijing Inhand Networks Technology Co., Ltd. | Method and system for detecting and locating single-phase ground fault on low current grounded power-distribution network |
| RU172271U1 (en) * | 2016-11-18 | 2017-07-03 | Федеральное государственное автономное образовательное учреждение высшего образования "Северо-Восточный федеральный университет имени М.К.Аммосова" | Installation for dynamic measurement of current-voltage characteristics of tunnel diodes |
| EP3460494A1 (en) * | 2017-09-26 | 2019-03-27 | Siemens Aktiengesellschaft | A method and apparatus for automatic detection of a fault type |
| US20190137556A1 (en) * | 2015-09-09 | 2019-05-09 | Beijing Inhand Networks Technology Co., Ltd. | Method and system for detecting and locating single-phase ground fault on low current grounded power-distribution network |
| CN110018389A (en) * | 2019-02-21 | 2019-07-16 | 国网山东省电力公司临沂供电公司 | A kind of transmission line of electricity on-line fault monitoring method and system |
| CN110879330A (en) * | 2019-12-02 | 2020-03-13 | 昆明理工大学 | Power distribution network single-phase earth fault development situation discrimination method based on zero sequence volt-ampere curve area |
| CN110988590A (en) * | 2019-11-25 | 2020-04-10 | 云南电网有限责任公司临沧供电局 | PCA-SVM model-based distribution network line selection method and system |
| CN111598166A (en) * | 2020-05-18 | 2020-08-28 | 国网山东省电力公司电力科学研究院 | Single-phase earth fault classification method and system based on principal component analysis and Softmax function |
| 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 |
| CN111796166A (en) * | 2020-08-27 | 2020-10-20 | 广东电网有限责任公司电力调度控制中心 | Power distribution network single-phase high-resistance earth fault line selection method, system and equipment |
| CN111812451A (en) * | 2020-06-04 | 2020-10-23 | 国电南瑞科技股份有限公司 | Phase current transient fault component-based distributed line selection method for power distribution network |
| CN112180210A (en) * | 2020-09-24 | 2021-01-05 | 华中科技大学 | Power distribution network single-phase earth fault line selection method and system |
| 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 |
| CN112255500A (en) * | 2020-10-12 | 2021-01-22 | 山东翰林科技有限公司 | Power distribution network weak characteristic fault identification method based on transfer learning |
| CN112462193A (en) * | 2020-11-05 | 2021-03-09 | 重庆邮电大学 | Power distribution network automatic reclosing judgment method based on real-time fault filtering data |
| CN113820624A (en) * | 2021-09-30 | 2021-12-21 | 南方电网科学研究院有限责任公司 | High-resistance grounding fault recognition device for power distribution network |
| 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 |
| CN114169231A (en) * | 2021-11-25 | 2022-03-11 | 上海交通大学 | Method for obtaining distribution line fault classification, positioning and line selection deep learning model based on transfer learning |
| CN114355240A (en) * | 2021-12-01 | 2022-04-15 | 国网安徽省电力有限公司电力科学研究院 | Power distribution network ground fault diagnosis method and device |
| CN114397531A (en) * | 2021-12-07 | 2022-04-26 | 北京智芯微电子科技有限公司 | Single-phase earth fault area positioning method and system, storage medium and feeder terminal |
| CN114414942A (en) * | 2022-01-14 | 2022-04-29 | 重庆大学 | Power transmission line fault identification classifier, identification method and system based on transient waveform image identification |
| WO2022183698A1 (en) * | 2021-03-05 | 2022-09-09 | 国网电力科学研究院武汉南瑞有限责任公司 | Artificial intelligence fault identification system and method based on power transmission line transient waveform |
-
2023
- 2023-02-15 CN CN202310119775.XA patent/CN116008731B/en active Active
Patent Citations (24)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2013024924A1 (en) * | 2011-08-16 | 2013-02-21 | 한국전력공사 | Apparatus and method for detecting a fault section of a power distribution system |
| US20160041216A1 (en) * | 2013-03-29 | 2016-02-11 | Beijing Inhand Networks Technology Co., Ltd. | Method and system for detecting and locating single-phase ground fault on low current grounded power-distribution network |
| US20150226780A1 (en) * | 2014-02-07 | 2015-08-13 | Mitsubishi Electric Research Laboratories, Inc. | Locating Multi-Phase Faults in Ungrounded Power Distribution Systems |
| US20190137556A1 (en) * | 2015-09-09 | 2019-05-09 | Beijing Inhand Networks Technology Co., Ltd. | Method and system for detecting and locating single-phase ground fault on low current grounded power-distribution network |
| RU172271U1 (en) * | 2016-11-18 | 2017-07-03 | Федеральное государственное автономное образовательное учреждение высшего образования "Северо-Восточный федеральный университет имени М.К.Аммосова" | Installation for dynamic measurement of current-voltage characteristics of tunnel diodes |
| EP3460494A1 (en) * | 2017-09-26 | 2019-03-27 | Siemens Aktiengesellschaft | A method and apparatus for automatic detection of a fault type |
| CN110018389A (en) * | 2019-02-21 | 2019-07-16 | 国网山东省电力公司临沂供电公司 | A kind of transmission line of electricity on-line fault monitoring method and system |
| 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 |
| CN110988590A (en) * | 2019-11-25 | 2020-04-10 | 云南电网有限责任公司临沧供电局 | PCA-SVM model-based distribution network line selection method and system |
| CN110879330A (en) * | 2019-12-02 | 2020-03-13 | 昆明理工大学 | Power distribution network single-phase earth fault development situation discrimination method based on zero sequence volt-ampere curve area |
| CN111598166A (en) * | 2020-05-18 | 2020-08-28 | 国网山东省电力公司电力科学研究院 | Single-phase earth fault classification method and system based on principal component analysis and Softmax function |
| CN111812451A (en) * | 2020-06-04 | 2020-10-23 | 国电南瑞科技股份有限公司 | Phase current transient fault component-based distributed line selection method for power distribution network |
| 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 |
| CN111796166A (en) * | 2020-08-27 | 2020-10-20 | 广东电网有限责任公司电力调度控制中心 | Power distribution network single-phase high-resistance earth fault line selection method, system and equipment |
| CN112180210A (en) * | 2020-09-24 | 2021-01-05 | 华中科技大学 | Power distribution network single-phase earth fault line selection method and system |
| CN112255500A (en) * | 2020-10-12 | 2021-01-22 | 山东翰林科技有限公司 | Power distribution network weak characteristic fault identification method based on transfer learning |
| CN112462193A (en) * | 2020-11-05 | 2021-03-09 | 重庆邮电大学 | Power distribution network automatic reclosing judgment method based on real-time fault filtering data |
| WO2022183698A1 (en) * | 2021-03-05 | 2022-09-09 | 国网电力科学研究院武汉南瑞有限责任公司 | Artificial intelligence fault identification system and method based on power transmission line transient waveform |
| CN113820624A (en) * | 2021-09-30 | 2021-12-21 | 南方电网科学研究院有限责任公司 | High-resistance grounding fault recognition device for power distribution network |
| 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 |
| CN114169231A (en) * | 2021-11-25 | 2022-03-11 | 上海交通大学 | Method for obtaining distribution line fault classification, positioning and line selection deep learning model based on transfer learning |
| CN114355240A (en) * | 2021-12-01 | 2022-04-15 | 国网安徽省电力有限公司电力科学研究院 | Power distribution network ground fault diagnosis method and device |
| CN114397531A (en) * | 2021-12-07 | 2022-04-26 | 北京智芯微电子科技有限公司 | Single-phase earth fault area positioning method and system, storage medium and feeder terminal |
| CN114414942A (en) * | 2022-01-14 | 2022-04-29 | 重庆大学 | Power transmission line fault identification classifier, identification method and system based on transient waveform image identification |
Non-Patent Citations (2)
| Title |
|---|
| ZHI-CE ZOU; JUN YAO: "Fault Protection of Urban Distribution Network Using a Superconducting Cable Joint", IEEE TRANSACTIONS ON APPLIED SUPERCONDUCTIVITY * |
| 陈振宁;李勇汇;彭辉;刘雯静;: "基于零序电压小波包能量比的配网单相高阻接地故障辨识分析", 科学技术与工程, no. 20 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116703284A (en) * | 2023-08-03 | 2023-09-05 | 八爪鱼人工智能科技(常熟)有限公司 | Fault identification method applied to refrigeration house management system and artificial intelligent server |
| CN116703284B (en) * | 2023-08-03 | 2023-10-17 | 八爪鱼人工智能科技(常熟)有限公司 | Fault identification method applied to refrigeration house management system and artificial intelligent server |
| CN117290756A (en) * | 2023-09-25 | 2023-12-26 | 重庆大学 | Power transmission line fault identification method and device based on federal learning and electronic equipment |
| CN117290756B (en) * | 2023-09-25 | 2024-04-16 | 重庆大学 | Power transmission line fault identification method and device based on federal learning and electronic equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116008731B (en) | 2023-08-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116008731B (en) | Power distribution network high-resistance fault identification method and device and electronic equipment | |
| Moradzadeh et al. | Turn-to-turn short circuit fault localization in transformer winding via image processing and deep learning method | |
| CN110082640B (en) | Distribution network single-phase earth fault identification method based on long-time memory network | |
| Perez et al. | An intelligent strategy for faults location in distribution networks with distributed generation | |
| CN111123048A (en) | Series fault arc detection device and method based on convolutional neural network | |
| WO2023083366A1 (en) | Method and apparatus for identifying running state of lightning arrester on basis of on-line monitoring data | |
| CN114414942A (en) | Power transmission line fault identification classifier, identification method and system based on transient waveform image identification | |
| CN113850330A (en) | Power distribution network fault cause detection method based on short-time Fourier transform and convolutional neural network | |
| Yu et al. | Faulty feeder detection of single phase-earth fault based on fuzzy measure fusion criterion for distribution networks | |
| CN115932484B (en) | Power transmission line fault identification and fault location method and device and electronic equipment | |
| De et al. | Recognition of impulse fault patterns in transformers using Kohonen's self-organizing feature map | |
| CN114580829A (en) | Power utilization safety sensing method, equipment and medium based on random forest algorithm | |
| CN112255500A (en) | Power distribution network weak characteristic fault identification method based on transfer learning | |
| CN113447766A (en) | Method, device, equipment and storage medium for detecting high-resistance ground fault | |
| Du et al. | Detection of single line-to-ground fault using convolutional neural network and task decomposition framework in distribution systems | |
| CN113780060A (en) | High-voltage switch cabinet situation sensing method based on multi-mode deep learning | |
| Baumann et al. | Impulse test fault diagnosis on power transformers using Kohonen's self-organizing neural network | |
| CN117272143A (en) | Power transmission line fault identification method and device based on gram angle field and residual error network | |
| CN110261746A (en) | Electric cable stoppage detection method based on oscillating wave voltage periodic attenuation characteristic | |
| Fahim et al. | An unsupervised protection scheme for overhead transmission line with emphasis on situations during line and source parameter variation | |
| Yuan et al. | Faulty feeder detection based on fully convolutional network and fault trust degree estimation in distribution networks | |
| CN107292512B (en) | Power equipment space-time multi-dimensional safety assessment method based on symbolic dynamics and hidden Markov model | |
| Ye et al. | Single pole‐to‐ground fault location method for mmc‐hvdc system using wavelet decomposition and dbn | |
| CN115879048A (en) | Series arc fault identification method and system based on WRFMDA model | |
| Wang et al. | The Cable Fault Diagnosis for XLPE Cable Based on 1DCNNs‐BiLSTM 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 | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20240617 Address after: 201206 building 17, zone B, jinlingzhidu, No. 2885, Shenjiang Road, Pudong New Area, Shanghai Patentee after: SHANGHAI ZEXIN ELECTRIC POWER TECHNOLOGY Co.,Ltd. Country or region after: China Address before: 400044 No. 174 Sha Jie street, Shapingba District, Chongqing Patentee before: Chongqing University Country or region before: China |