CN115932484A - Method and device for identifying and ranging faults of power transmission line and electronic equipment - Google Patents
Method and device for identifying and ranging faults of power transmission line and electronic equipment Download PDFInfo
- Publication number
- CN115932484A CN115932484A CN202310114821.7A CN202310114821A CN115932484A CN 115932484 A CN115932484 A CN 115932484A CN 202310114821 A CN202310114821 A CN 202310114821A CN 115932484 A CN115932484 A CN 115932484A
- Authority
- CN
- China
- Prior art keywords
- fault
- volt
- characteristic curve
- ampere characteristic
- transmission line
- 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
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
Abstract
The invention relates to a method for identifying and ranging faults of a power transmission line, which comprises the following steps: acquiring line data of the power transmission line; establishing a volt-ampere characteristic curve of the power transmission line based on the line data; normalizing the volt-ampere characteristic curve into a fault type identification volt-ampere characteristic curve graph; identifying the fault type identification volt-ampere characteristic curve according to a fault type identification model which is constructed in advance; constructing a fault reason and distance identification volt-ampere characteristic curve graph corresponding to a fault phase based on the identified fault type; and identifying the fault reason and distance identification volt-ampere characteristic curve according to a pre-constructed fault reason and distance identification model so as to output the identified fault reason and fault distance. The invention adopts the volt-ampere characteristic diagram as input, can realize the self-extraction of characteristic quantity without depending on a signal processing algorithm, can effectively solve the problem that the recording data contains white noise, and has low requirements on the sampling frequency of voltage and current.
Description
Technical Field
The invention relates to the technical field of power transmission line fault identification, in particular to a power transmission line fault identification and fault location method, a power transmission line fault identification and fault location device and electronic equipment based on volt-ampere characteristic curve image identification.
Background
Safe and reliable operation of the power transmission line is an important condition for ensuring safe and stable operation of the power system. Overhead transmission lines are widely distributed, the structure is more complex, the operating environment is changeable, and various types of faults are easy to occur to the transmission lines under the influence of severe weather conditions, natural disasters, artificial damage and the like. The fault type and the reason can be timely and accurately identified, and the method has important significance for guiding self-adaptive reclosing and recovering the power transmission of the line, reducing the shutdown time of the line and ensuring the safe and stable operation of a power system. The improvement of the primary power equipment sensing and measuring technology improves the data sampling precision, enriches the fault information in the fault recording data and provides possibility for sensing the running state of the power transmission line.
In terms of constructing a fault type identifier, the existing research methods generally use a fault type sample set with uniformly distributed parameters such as fault location, type, starting phase angle, transition resistance, and the like to establish the fault type identifier. However, from the fault trip records obtained by the power grid company, the number of samples has a serious class imbalance problem due to fault types such as single-phase grounding, two-phase short circuit and the like, and fault causes such as lightning stroke, mountain fire, foreign matters and the like. The problem of small failure sample sets due to lack of serviceability is also not fully considered. In addition, the existing research method generally adopts the method of respectively establishing a fault type identifier and a fault distance identifier to realize fault identification and fault location, and does not consider the incidence relation between the fault location and the fault identification.
In addition, in the aspect of selecting an input set, the existing research method generally depends on signal analysis and characteristic quantity extraction, the selected fault characteristics are easily influenced by factors such as voltage and current waveforms, fault distance and transition resistance, the characteristic selection process is complex, and the classification and identification accuracy is not ideal. Also, many of these schemes require high sampling rates. However, the sampling frequency of the Current Transformer (CT) in practical situations often does not exceed 20kHz, which is also a reason why such methods are not effective in practice.
Therefore, it is necessary to design a new method for identifying faults of the transmission line to solve the above technical problems.
Disclosure of Invention
In view of this, the invention provides a method for identifying and ranging a fault of a power transmission line, which uses a volt-ampere characteristic diagram as an input, does not depend on a signal processing algorithm, can realize self-extraction of a characteristic quantity, can effectively solve the problem that recording data contains white noise, and has low requirements on sampling frequency of voltage and current. In addition, the method is carried out on the basis of a deep learning neural network model trained by adopting volt-ampere characteristic curve images, so that the interference of fault distance to fault reasons and the interference of the fault reasons to fault distance measurement can be reduced, and the identification of the fault reasons and the fault distance is realized.
The invention discloses a method for identifying and ranging faults of a power transmission line in a first aspect, which comprises the following steps:
acquiring line data of the power transmission line;
establishing a volt-ampere characteristic curve of the power transmission line based on the acquired line data of the power transmission line;
normalizing a volt-ampere characteristic curve constructed based on line data of a power transmission line into a single total volt-ampere characteristic curve image as a fault type identification volt-ampere characteristic curve graph;
identifying the fault type identification volt-ampere characteristic curve according to a fault type identification model which is constructed in advance, and outputting the identified fault type;
acquiring a fault phase corresponding to the fault type of the power transmission line based on the identified fault type, and then constructing a single fault phase volt-ampere characteristic curve image corresponding to the fault phase as a fault reason and distance identification volt-ampere characteristic curve;
and identifying and processing the fault reason and distance identification volt-ampere characteristic curve according to a pre-constructed fault reason and distance identification model so as to output the identified fault reason and fault distance.
In the pre-constructed fault type identification model, the fault type mark comprises: single phase grounding, two-phase short circuit grounding, three-phase short circuit. In the implementation process, the line data of the power transmission line to be tested is acquired firstly, then the acquired current line data is converted into corresponding volt-ampere characteristic curves, and the volt-ampere characteristic curves can be multiple according to the positions and the phase differences of the line data. Normalizing the obtained multiple volt-ampere characteristic curves into a single total volt-ampere characteristic curve image, inputting the single total volt-ampere characteristic curve image serving as a fault type identification volt-ampere characteristic curve graph into a pre-constructed fault type identification model, identifying through the pre-constructed fault type identification model, and finally outputting an identified fault type mark. When the fault type of the corresponding tested power transmission line is identified, determining a fault phase corresponding to the power transmission line according to the fault type, then constructing a single fault phase volt-ampere characteristic curve image corresponding to the fault phase based on the determined fault phase, then inputting the obtained single fault phase volt-ampere characteristic curve image serving as a fault reason and distance identification volt-ampere characteristic curve into a pre-constructed fault reason and distance identification model, and finally obtaining the identified fault reason and fault distance. The "single total volt-ampere characteristic curve image" refers to a single image including all volt-ampere characteristic curves constructed based on line data of the power transmission line, and the "single fault phase volt-ampere characteristic curve image" refers to a single image including only the volt-ampere characteristic curve corresponding to the fault phase. That is, compared to a single total voltammogram image, a single faulty phase voltammogram image does not include voltammograms corresponding to non-faulty phases.
In the process, the volt-ampere characteristic curve graph is used as input, a signal processing algorithm is not relied on, the self-extraction of the characteristic quantity can be realized, the problem that the recorded wave data contains white noise can be effectively solved, and the requirement on sampling frequency is low.
Further, in the method for identifying and ranging the fault of the power transmission line, a single fault phase volt-ampere characteristic curve image corresponding to the fault phase is constructed, and the method comprises the following steps: and processing the single total volt-ampere characteristic curve image based on the obtained fault phase corresponding to the fault type, and obtaining a single fault phase volt-ampere characteristic curve image corresponding to the fault phase. In this embodiment, a single faulty phase voltammetry characteristic image is obtained by directly processing a single previously obtained total voltammetry characteristic image, that is, a voltammetry characteristic corresponding to a faulty phase is directly obtained from the single previously obtained total voltammetry characteristic image to form the single faulty phase voltammetry characteristic image.
In other embodiments, a single faulty phase current-voltage characteristic curve image can be obtained by: based on the judged fault phase corresponding to the fault type of the power transmission line, a volt-ampere characteristic curve corresponding to the fault phase is directly screened out from a volt-ampere characteristic curve constructed based on line data of the power transmission line, and then the screened volt-ampere characteristic curve corresponding to the fault phase is subjected to normalization processing to form a single fault phase volt-ampere characteristic curve image.
Further, in the method for identifying and ranging the fault of the power transmission line, the line data includes wave recording data of secondary side three-phase voltage and current signals at two ends of the power transmission line and wave recording data of zero-sequence voltage and current signals; the establishing of the volt-ampere characteristic curve of the power transmission line based on the acquired line data of the power transmission line comprises the following steps: and constructing a three-phase volt-ampere characteristic curve and a zero-sequence volt-ampere characteristic curve at the head end side of the power transmission line, and constructing a three-phase volt-ampere characteristic curve and a zero-sequence volt-ampere characteristic curve at the tail end side of the power transmission line.
The three phases of the transmission line are phase A, phase B and phase C. In the implementation process, the acquired line data includes: transient recording data of a secondary side A phase voltage and current of a head end (M end) of the power transmission line, transient recording data of a secondary side B phase voltage and current of the head end of the power transmission line, transient recording data of a secondary side C phase voltage and current of the head end of the power transmission line, transient recording data of a secondary side zero sequence voltage and current of the head end of the power transmission line, transient recording data of a secondary side A phase voltage and current of a tail end (N end) of the power transmission line, transient recording data of a secondary side B phase voltage and current of a tail end of the power transmission line, transient recording data of a secondary side C phase voltage and current of the tail end of the power transmission line, and transient recording data of a secondary side zero sequence voltage and current of the tail end of the power transmission line. And corresponding to the data one by one, constructing six volt-ampere characteristic curves of the measured power transmission line, and normalizing the six volt-ampere characteristic curves into a single total volt-ampere characteristic curve image serving as a fault type identification volt-ampere characteristic curve graph for fault type identification. After the fault type is identified, a single fault phase volt-ampere characteristic curve image only comprising a volt-ampere characteristic curve corresponding to the phase related to the fault type is constructed based on the phase related to the fault type, and then the single fault phase volt-ampere characteristic curve image is used as a fault cause and distance identification chart and is input into a fault cause and distance identification model for identification.
Further, in the method for identifying and ranging a fault of a power transmission line, before obtaining line data of the power transmission line, the method further includes:
constructing a fault type original identification model and a fault reason and distance original identification model;
acquiring line data of a plurality of fault transmission lines for fault samples to form a fault sample data set;
acquiring a line fault type sample set and a line fault reason and distance sample set based on the fault sample data set;
inputting the obtained line fault sample set into the original fault type identification model to perform transfer learning training until convergence, so as to obtain the fault type identification model; and (c) a second step of,
and inputting the obtained line fault reason and distance sample set into the fault reason and distance original identification model to perform transfer learning training until convergence, so as to obtain the fault reason and distance identification model.
Further, in the method for identifying and ranging faults of a power transmission line, the obtaining a line fault type sample set includes:
constructing a volt-ampere characteristic curve of a corresponding fault power transmission line based on the acquired fault sample data set;
normalizing the constructed volt-ampere characteristic curve of the corresponding fault power transmission line into a single total volt-ampere characteristic curve image sample as a fault type identification volt-ampere characteristic curve sample;
a plurality of samples of the fault type identifying voltammogram are acquired to form a sample set of line fault types.
Further, the acquiring the line fault cause and the distance sample set includes:
processing the single total volt-ampere characteristic curve image sample based on a fault type mark in a fault sample data set, and obtaining a single fault phase volt-ampere characteristic curve image sample corresponding to a fault phase;
and acquiring a plurality of single fault phase volt-ampere characteristic curve image samples to form a line fault reason and distance sample set.
In the implementation process, the line fault type sample set or the line fault reason and distance sample set is constructed based on the volt-ampere characteristic curve constructed by the sampling data. That is, the training and testing of the corresponding original model are performed based on the voltammogram image.
In the prior art, fault type identifiers and fault distance identifiers are respectively established to realize fault identification and fault distance measurement. In the invention, because the volt-ampere characteristic curve image is used as the input characteristic value of the model to train the deep learning neural network, the training and the testing of the fault reason and the fault distance identification in the same deep learning neural network model can be realized by means of the known data of the fault sample and the parameter characteristic of the volt-ampere characteristic curve, and the interference of the fault distance to the fault reason and the interference of the fault reason to the fault distance measurement can be reduced, so that the identification of the fault reason and the fault distance can be simultaneously realized by means of the fault reason and the fault distance model. Specifically, in the method, the fault cause and distance identification model finally formed after convergence can simultaneously extract the fault cause and the characteristics of the distance volt-ampere characteristic curve image, such as harmonic waves, amplitude, slope and the like, the characteristic graph is formed through the add layer of the model, the characteristics influencing fault cause identification are automatically extracted from the characteristic graph through the full connection layer (fc _ 3), and the characteristics influencing fault distance identification are automatically extracted from the characteristic graph through the full connection layer (fc _ 2).
Further, in the method for identifying and ranging the fault of the power transmission line, the line data of the faulty power transmission line includes: the method comprises the steps of recording data of secondary side three-phase voltage and current signals at two ends of a power transmission line, recording data of secondary side zero sequence voltage and current signals, a fault type mark, a fault reason mark and a fault distance mark.
Further, in the method for identifying and ranging a fault of a power transmission line, in the step of obtaining line data of a plurality of faulty power transmission lines for fault samples to form a fault sample data set, the method includes:
taking the proportion of each fault type to the total fault type and the proportion of each fault reason to the total fault reason as classification weights; and/or, in the step of obtaining line data of a plurality of faulty transmission lines for the faulty sample to form a faulty sample data set, the method further includes: a semi-supervised model is established to label unlabeled fault samples.
By referring to the statistical probability distribution of the fault types, the fault type phases and the fault reasons in the actual fault records, the classification weight of a deep learning neural network model (an original recognition model) is set, the training of the fault recognition model is facilitated, and the generalization capability and the application effect of the fault recognition model to the actual situation can be improved. In addition, by referring to the situations that the labeled samples are absent and a large number of unlabeled samples exist in actual situations, a re-labeling method based on semi-supervised learning is established, so that a real sample data set can be effectively expanded, and the training of a fault identification original model is facilitated.
Further, in the method for identifying and ranging the fault of the power transmission line, the original identification model of the fault type is a single-output convolutional neural network architecture, and the original identification model of the fault cause and the distance is a multi-output convolutional neural network architecture.
The second aspect of the invention also discloses a transmission line fault identification and fault location device, which comprises:
the first acquisition unit is used for acquiring line data of the power transmission line;
the volt-ampere characteristic curve construction unit is used for constructing a volt-ampere characteristic curve of the power transmission line based on the acquired line data of the power transmission line;
the normalization processing unit is used for normalizing the volt-ampere characteristic curve constructed based on the line data of the power transmission line into a single total volt-ampere characteristic curve image which is used as a fault type identification volt-ampere characteristic curve graph;
the fault type identification unit is used for identifying the fault type identification volt-ampere characteristic curve according to a pre-constructed fault type identification model and outputting the identified fault type;
the image construction unit is used for obtaining a fault phase corresponding to the fault type of the power transmission line based on the identified fault type, and then constructing a single fault phase volt-ampere characteristic curve image corresponding to the fault phase as a fault reason and distance identification volt-ampere characteristic curve graph;
and the fault reason and distance identification unit is used for identifying and processing the fault reason and distance identification volt-ampere characteristic curve graph according to a pre-constructed fault reason and distance identification model so as to output the identified fault reason and fault distance.
Further, in the apparatus disclosed in the second aspect of the present invention, the apparatus further includes:
the model construction unit is used for respectively constructing a fault type original identification model and a fault reason and distance original identification model before acquiring the line data of the power transmission line;
the second acquisition unit is used for acquiring line data of a plurality of fault transmission lines for fault samples to form a fault sample data set;
the system comprises a sample set classification unit, a fault analysis unit and a fault analysis unit, wherein the sample set classification unit is used for acquiring a line fault type sample set and a line fault reason and distance sample set based on the fault sample data set;
the first training unit is used for inputting the obtained line fault sample set into the fault type original identification model to carry out transfer learning training until convergence, so as to obtain the fault type identification model; and the number of the first and second groups,
and the second training unit is used for inputting the obtained line fault reason and distance sample set into the original fault reason and distance identification model to perform transfer learning training until convergence, so as to obtain the fault reason and distance identification model.
In a third aspect of the present invention, an electronic device is disclosed, which includes a memory for storing a computer program and a processor for operating the computer program to make the electronic device execute the method for identifying and ranging the power transmission line fault described in the first aspect.
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 and ranging faults of the power transmission line described in the first aspect of the present invention is executed.
Has the advantages that: according to the method for identifying and locating the faults of the power transmission line, the volt-ampere characteristic diagram is used as input, a signal processing algorithm is not relied on, the self-extraction of characteristic quantity can be realized, the problem that recorded wave data contains white noise can be effectively solved, and the requirements on the sampling frequency of voltage and current are low. In addition, in the invention, because the volt-ampere characteristic curve is used as input, the fault reason and the fault distance can be respectively identified in the same model, and the technical problem of mutual interference between the fault reason and the fault distance in the prior art is avoided.
The method for identifying and ranging the transmission line fault is disclosed in detail below by combining the embodiments shown in the attached drawings and the reference numerals.
Drawings
Fig. 1 is a flowchart illustrating steps of a method for identifying and ranging a fault of a power transmission line according to an embodiment of the present invention.
FIG. 2 is a flow chart illustrating the steps of pre-building and training an original recognition model in an embodiment of the present invention.
Fig. 3 shows an example of a single total voltammetry curve image generated by normalizing transient oscillometric data in an embodiment of the present invention.
Fig. 4 shows an example of a single faulty phase current-voltage characteristic curve image in an embodiment of the present invention.
FIG. 5 shows a schematic block diagram of a single output convolutional neural network in an embodiment of the present invention.
FIG. 6 shows a schematic block diagram of a multi-output convolutional neural network in an embodiment of the present invention.
7 (a) - (c) show test result diagrams of the deep learning neural network model based on the imbalance-like fault sample in the embodiment of the invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be described clearly and completely with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments of the present invention without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1, an embodiment of the present invention discloses a method for identifying and ranging a fault of a power transmission line, including the following steps:
acquiring line data of the power transmission line;
establishing a volt-ampere characteristic curve of the power transmission line based on the acquired line data of the power transmission line;
normalizing a volt-ampere characteristic curve constructed based on line data of a power transmission line into a single total volt-ampere characteristic curve image as a fault type identification volt-ampere characteristic curve graph;
identifying the fault type identification volt-ampere characteristic curve according to a pre-constructed fault type identification model, and outputting the identified fault type;
acquiring a fault phase corresponding to the fault type of the power transmission line based on the identified fault type, and then constructing a single fault phase volt-ampere characteristic curve image corresponding to the fault phase as a fault reason and a distance identification volt-ampere characteristic curve;
and identifying and processing the fault reason and distance identification volt-ampere characteristic curve graph according to a pre-constructed fault reason and distance identification model so as to output the identified fault reason and fault distance.
In the above embodiment, constructing a single fault phase current-voltage characteristic curve image corresponding to a fault phase includes: and processing the single total volt-ampere characteristic curve image based on the obtained fault phase corresponding to the fault type, and obtaining a single fault phase volt-ampere characteristic curve image corresponding to the fault phase. In this embodiment, a single faulty phase voltammetry characteristic image is obtained by directly processing a single previously obtained total voltammetry characteristic image, that is, a voltammetry characteristic corresponding to a faulty phase is directly obtained from the single previously obtained total voltammetry characteristic image to form the single faulty phase voltammetry characteristic image.
In the above embodiment, the line data includes wave recording data of secondary side three-phase voltage and current signals at two ends of the power transmission line, and wave recording data of zero sequence voltage and current signals; the establishing of the volt-ampere characteristic curve of the power transmission line based on the acquired line data of the power transmission line comprises the following steps: and constructing a three-phase volt-ampere characteristic curve and a zero sequence volt-ampere characteristic curve at the head end side of the power transmission line, and constructing a three-phase volt-ampere characteristic curve and a zero sequence volt-ampere characteristic curve at the tail end side of the power transmission line. Preferably, when the transmission line fault is actually identified through the constructed neural network model, firstly, a secondary side signal of rated voltage of the transmission line is taken as a normalized denominator of a voltage transformer signal, and a secondary side signal of 5 times of rated current is taken as a normalized denominator of a current transformer, so that the secondary side signal is subjected to normalization processing. Then, a fault type identification volt-ampere characteristic curve graph, a fault cause identification volt-ampere characteristic curve graph and a distance identification volt-ampere characteristic curve graph are constructed by taking the normalized voltage as an abscissa and the normalized current as an ordinate, as shown in fig. 3 and 4.
Preferably, the fault type identification volt-ampere characteristic curve graph is composed of a plurality of volt-ampere characteristic curves of 4 rows and 2 columns, wherein the 1 st row 1 is a volt-ampere characteristic curve graph of an a phase at an M end of the power transmission line, the 1 st row 2 is a volt-ampere characteristic curve graph of an a phase at an N end of the power transmission line, the 2 nd row 1 is a volt-ampere characteristic curve graph of a B phase at the M end of the power transmission line, the 2 nd row 2 is a volt-ampere characteristic curve graph of an B phase at the N end of the power transmission line, the 3 rd row 1 is a volt-ampere characteristic curve graph of an M phase at the M end of the power transmission line, the 3 rd row 2 is a volt-ampere characteristic curve graph of a C phase at the N end of the power transmission line, the 4 th row 1 is a volt-ampere characteristic curve graph of a zero sequence volt-ampere characteristic curve at the M end of the power transmission line, and the 4 th row 2 is a zero sequence volt-ampere characteristic curve graph of a zero sequence at the N end of the power transmission line. As shown in fig. 3.
Preferably, the volt-ampere characteristic curve for identifying the fault cause and the distance is composed of two volt-ampere characteristic curves in 1 row and 2 columns, in this embodiment, the fault phase relates to a phase a of the power transmission line, where the 1 st row and 1 column are volt-ampere characteristic curves of a phase a at an M end of the power transmission line, and the 1 st row and 2 column are volt-ampere characteristic curves of a phase a at an N end of the power transmission line. In the invention, if the fault type is a fault related to the A phase, the A and B phases, the A and C phases or the ABC three phases of the power transmission line, only a single fault phase volt-ampere characteristic curve graph containing the volt-ampere characteristic curve corresponding to the A phase is constructed; if the fault type is a fault related to two phases B, B and C of the power transmission line, only constructing a single fault phase volt-ampere characteristic curve graph containing a volt-ampere characteristic curve corresponding to the phase B; if the fault type is a fault involving only phase C of the transmission line, then only a single fault phase voltage current characteristic curve containing voltage current characteristic curves corresponding to phase C is constructed.
The method for identifying and ranging the faults of the power transmission line also provides a method for constructing an identification model in advance, which is implemented before the step of acquiring the line data of the power transmission line of the method, and is shown in a combined manner in figure 2, wherein the process comprises the following steps:
constructing a fault type original identification model and a fault reason and distance original identification model;
acquiring line data of a plurality of fault transmission lines for generating fault samples to form a fault sample data set;
acquiring a line fault type sample set and a line fault reason and distance sample set based on the fault sample data set;
inputting the obtained line fault sample set into the original fault type identification model to perform transfer learning training until convergence, so as to obtain the fault type identification model; and (c) a second step of,
and inputting the obtained line fault reason and distance sample set into the original fault reason and distance identification model to perform transfer learning training until convergence, so as to obtain the fault reason and distance identification model.
In the above embodiment, obtaining a line fault type sample set includes:
constructing a volt-ampere characteristic curve of a corresponding fault power transmission line based on the acquired fault sample data set;
normalizing the constructed volt-ampere characteristic curve of the corresponding fault power transmission line into a single total volt-ampere characteristic curve image sample as a fault type identification volt-ampere characteristic curve sample;
a plurality of samples of the fault type identifying current-voltage characteristic plot are acquired to form a line fault type sample set.
Similarly, in the specific implementation process of constructing the fault type sample set, the invention firstly uses the secondary side signal of the rated voltage of the power transmission line as the normalized denominator of the signal of the voltage transformer, and uses the secondary side signal of 5 times of the rated current as the normalized denominator of the current transformer, and the normalization processing is carried out on the secondary side signal. Then, using the normalized voltage as an abscissa and the normalized current as an ordinate, a fault type identification voltammogram sample, a fault cause and distance identification voltammogram sample are constructed, as shown in fig. 3 and 4.
In addition, in the specific implementation process of constructing the fault type sample set, the fault type identification volt-ampere characteristic curve sample consists of a plurality of volt-ampere characteristic curves in 4 rows and 2 columns, wherein the 1 st row 1 is a volt-ampere characteristic curve of a phase a at the M end of the power transmission line, the 1 st row 2 is a volt-ampere characteristic curve of a phase a at the N end of the power transmission line, the 2 nd row 1 is a volt-ampere characteristic curve of a phase B at the M end of the power transmission line, the 2 nd row 2 is a volt-ampere characteristic curve of a phase B at the N end of the power transmission line, the 3 rd row 1 is a volt-ampere characteristic curve of a phase C at the M end of the power transmission line, the 3 rd row 2 is a volt-ampere characteristic curve of a phase C at the N end of the power transmission line, the 4 th row 1 is a volt-ampere characteristic curve of a zero sequence at the M end of the power transmission line, and the 4 th row 2 is a volt-ampere characteristic curve of a zero sequence at the N end of the power transmission line. As shown in fig. 3.
In addition, similarly, in the specific implementation process of constructing the fault type sample set, the fault cause and distance identification volt-ampere characteristic curve sample consists of two volt-ampere characteristic curves in 1 row and 2 columns, in this embodiment, the fault phase relates to the a phase of the power transmission line as an example, wherein the 1 st row and 1 column are volt-ampere characteristic curves of the M-end and a-phase of the power transmission line, and the 1 st row and 2 column are volt-ampere characteristic curves of the N-end and a-phase of the power transmission line. In the specific implementation process of constructing the fault type sample set, if the fault type marked by the sample is a fault related to the A phase, the A and B phases, the A and C phases or the ABC three phases of the power transmission line, only a single fault phase volt-ampere characteristic curve sample containing a volt-ampere characteristic curve corresponding to the A phase is constructed; if the fault type marked by the sample is a fault related to the B phase, the B phase and the C phase of the power transmission line, only constructing a single fault phase volt-ampere characteristic curve graph sample containing a volt-ampere characteristic curve corresponding to the B phase; if the type of fault marked by the sample is a fault only involving the C phase of the transmission line, only a single sample of the voltage-current characteristic curve of the faulted phase is constructed, which contains the voltage-current characteristic curve corresponding to the C phase.
In the above embodiment, the obtaining the line fault cause and the distance sample set includes:
processing the single total volt-ampere characteristic curve image sample based on a fault type mark in a fault sample data set, and obtaining a single fault phase volt-ampere characteristic curve image sample corresponding to a fault phase;
and acquiring a plurality of single fault phase volt-ampere characteristic curve image samples to form a line fault reason and distance sample set.
In the above embodiment, the line data of the faulty transmission line includes: the method comprises the steps of recording data of secondary side three-phase voltage and current signals at two ends of a power transmission line, recording data of secondary side zero sequence voltage and current signals, a fault type mark, a fault reason mark and a fault distance mark.
In the above embodiment, in the step of obtaining line data of a plurality of faulty transmission lines used for a faulty sample to form a faulty sample data set, the method includes: and taking the proportion of each fault type to the total fault type and the proportion of each fault reason to the total fault reason as classification weights. The method specifically comprises the following steps: and counting the proportion of each fault type sample in the fault type sample set, counting the proportion of each fault reason sample in the fault reason sample set, and taking the proportion of each fault type in the total fault type and the proportion of each fault reason in the total fault reason as classification weights.
In the invention, the generation of the fault sample can be realized by a sample generation method in the prior art, for example, the fault sample can be generated by collecting fault labels of the transmission lines of an actual distribution network in a certain area, or a simulation model can be constructed according to actual transmission line sections under the condition that a large number of effective fault class labels are lacked in actual work, so that a fault sample set with unbalanced fault types is generated for model training. The invention carries out statistical analysis on the fault samples of the power grid in a certain area, and the distribution condition of the fault types is found as follows: 90.66% of single-phase grounding short circuit, 4.8% of two-phase short circuit, 3.8% of two-phase grounding short circuit and 0.73% of three-phase short circuit. Further analysis finds that except for single-phase grounding short circuit, the transition resistance of other interphase faults is mostly electric arc, and the transition resistance is small; the reason for single-phase ground short circuit is more, and the problem of unbalanced distribution also exists. According to statistics, lightning strike, foreign matters, mountain fire and trees are main reasons causing the faults of the high-voltage overhead line. Wherein, the lightning stroke fault accounts for 44.33%, the mountain fire fault accounts for 28.08%, the foreign matter fault accounts for 4.43%, and the tree flash fault accounts for 2.96%. And lightning stroke faults do not occur uniformly, and in practical situations, a lightning stroke line and a lightning stroke tower respectively account for 10 percent and 90 percent. The fault phase of the transmission line also shows class imbalance. Because the transmission lines are mostly arranged in a triangular or horizontal arrangement, the B-phase conductors are usually between the A-phase conductors and the C-phase conductors. Therefore, in a single-phase grounding short circuit caused by lightning stroke, trees, foreign matters and mountain fire, the number of B-phase wires is less than that of A-phase wires and C-phase wires. In addition, the invention also shows certain class imbalance according to statistics of the geographical distribution of the fault positions along the power transmission line, and certain places are easy to have faults due to special terrains and microclimates, but the fault positions are relatively uniform as a whole. The power equipment is easy to discharge at the voltage peak value due to self aging or local defects, but for the transmission line, the occurrence time of the fault is mostly determined by external factors, so the distribution of the fault phase angle is relatively uniform.
According to the distribution characteristics, the invention sets 19 types of classification marks as follows: the lightning protection system comprises an A-phase lightning stroke line ground short circuit fault, a B-phase lightning stroke line ground short circuit fault, a C-phase lightning stroke line ground short circuit fault, an A-phase lightning stroke tower ground short circuit fault, a B-phase lightning stroke tower ground short circuit fault, a C-phase lightning stroke tower ground short circuit fault, an A-phase foreign object ground short circuit fault, a B-phase foreign object ground short circuit fault, a C-phase tree flash ground short circuit fault, an AB two-phase fault, a BC two-phase fault, an AC two-phase fault, an AB two-phase ground short circuit fault, a BC two-phase ground short circuit fault, an AC two-phase ground short circuit fault and a three-phase short circuit fault. And sets classification weights according to equations (1) and (2).
weight＝[w 1 ,w 2 ,...,w 19 ] (1)
Wherein w i ,i∈[1,19]Indicating the number of the ith classification mark; sum () vector summation; w i ,i∈[1,19]Representing the weight of the ith class label.
In the above embodiment, in the step of obtaining line data of a plurality of faulty transmission lines used for a faulty sample to form a faulty sample data set, the method further includes: a semi-supervised model is established to label unlabeled fault samples. Wherein the step of establishing a semi-supervised model to label the unlabeled exemplars comprises:
normalizing voltage and current signals at two ends of the power transmission line;
according to the fault sample with or without fault reason mark, dividing the sample into a marked sample set and an unmarked sample set;
and inputting the unmarked sample set and the marked sample set and the marks into a semi-supervised learning model to be marked again.
Specifically, in the step, a K-neighbor neural network is used as a semi-supervised learning model, and voltage and current discrete signals of a power transmission line are used as input, so that the unlabeled samples are re-labeled. Mutual k-nearest neighbor graph (MKG) establishes a local neighborhood relationship model between labeled and unlabeled data through the similarity graph as an undirected graph. The nodes in the graph represent observations, and the undirected edges represent connections between observations. If the distance between a node and a node is non-zero or greater than a certain threshold, the similarity graph connects two nodes with an edge. Edge pair similarity S between two nodes i，j Weighting, S i，j The calculation of (2) is as shown in the formula (3).
Where σ is a particular kernel size.
In specific implementation, an n-by-K matrix F (0) is initialized, wherein n is the number of nodes and K is the number of classes. The first l rows correspond to marked points. Each row contains a 1 in the column of the real class label corresponding to the point and a 0 in each of the other columns. The last u rows correspond to unmarked points and contain one 0 in all columns. Then, starting from t =1, the F matrix is updated using a probability transition matrix P, such that F (t) = PF (t-1), where P i，j Is the probability that node i transmits the label information to node j, which is calculated as equation (4).
The iteration aims to ensure that the true class label of the first l lines is not lost, i.e. the first l lines of F (t) need to be kept equal to the initial value in F (0). Finally, through multiple iterations, when the F values converge, the column in F with the largest score per row corresponds to the fitted class label.
In the above embodiment, the original identification model of the fault type is a single-output convolutional neural network architecture, and the original identification model of the fault cause and distance is a multi-output convolutional neural network architecture. Preferably, the original recognition model can employ a pre-trained AlexNet network model.
The original fault type identification model is a single-output convolutional neural network architecture based on a residual error neural network, and the structure of the original fault type identification model is shown in fig. 5, and the original fault type identification model comprises convolutional layers conv _1, the size of the convolutional layers conv _1 is 5 multiplied by 1, and the sliding step length of the convolutional layers conv _1 is 1; convolution layer conv _2, with size of 3 × 32 × 1 and sliding step length of 2; convolution layer conv _3 with size of 3 × 32 × 1 and sliding step length of 1; convolution layer conv _ skip, size 1 × 32 × 1, sliding step length 2; normalization layers backnorm _1, backnorm _2, backnorm _3 and backnorm _ skip; activating function layers relu _1, relu _2, relu _3 and relu _ skip; full connection layer fc _3, size 1000; full connection layer fc, size 19; dropout layer, size 0.5.
In the invention, the size of the sample of the fault type identification volt-ampere characteristic diagram is adjusted to 224 x 3, then the sample set of the fault type identification volt-ampere characteristic diagram is input into a fault type identifier, the characteristic quantities of the fault type identification volt-ampere characteristic diagram are extracted through a convolution layer conv _1, a normalization layer backnorm _1, an activation function layer relu _1, a convolution layer conv _2, a normalization layer backnorm _2, a convolution layer conv _3, a normalization layer backnorm _3 and an activation function layer relu _3, and the characteristic quantities of the convolution layer conv _ skip, the normalization layer thrum _ skip and the activation function layer relu _ skip in a residual branch are fused. And finally, classifying the fault types by the fused characteristic quantity through a fully-connected neural network.
In the training and testing of the fault type original recognition model, 70% of a fault type recognition sample set is randomly extracted to serve as a training sample set, the rest is used as a testing sample set, the training sample set serves as target domain data and is input to a single-output convolutional neural network model, the fault type is used as output for training, the maximum training frequency is set, whether convergence occurs is judged, the training of the model is completed after the maximum training frequency is reached on the premise of convergence, and if the training does not converge, the maximum training frequency is increased until the convergence occurs. And finally, inputting the accuracy of the test sample set calculation model to obtain a final classification model.
After determining the fault type, establishing a fault cause and distance original identification model based on a residual error neural network, which is a multi-output convolutional neural network, and the structure of which is shown in fig. 6, includes: convolution layer conv _1, size 5 × 5 × 1, sliding step length 1; convolution layer conv _2, with size of 3 × 32 × 1 and sliding step length of 2; convolution layer conv _3, with size of 3 × 32 × 1 and sliding step length of 1; convolution layer conv _ skip, size 1 × 32 × 1, sliding step length 2; normalization layers backnorm _1, backnorm _2, backnorm _3 and backnorm _ skip; activating function layers relu _1, relu _2, relu _3 and relu _ skip; a full connection layer fc _3 with a size of 1000; full connection layer fc, size 19; a full connection layer fc _2 of size 1; dropout layer, size 0.5.
The method inputs a fault cause and distance identification volt-ampere characteristic curve sample set into a fault cause and distance original identification model for training, and extracts the fault cause and the characteristic quantity of the distance identification volt-ampere characteristic curve through a convolution layer conv _1, a normalization layer backnorm _1, an activation function layer relu _1, a convolution layer conv _2, a normalization layer backnorm _2, an activation function layer relu _2, a convolution layer conv _3, a normalization layer backnorm _3 and an activation function layer relu _3, and fuses the characteristic quantities of a residual branch convolution layer conv _ skip, a normalization layer backnorm _ skip and an activation function layer relu _ skip. Finally, identifying the fault reason through the double fully-connected layers, the dropout layer and the softmax layer, and predicting the fault distance through the fully-connected layer fc _ 2.
In the training process of the original identification model of the fault reason and the distance, 70% of the identification sample set of the fault reason is randomly extracted to serve as a training sample set, the rest is used as a test sample set and is input to the multi-output convolutional neural network model as target domain data, the fault reason and the fault distance are used as outputs to carry out training, the maximum training frequency is set and whether convergence occurs is judged, the training of the model is completed after the maximum training frequency is reached on the premise of convergence, and if the training does not converge, the maximum training frequency is increased until the convergence. The fault reason and distance identification model finally formed after convergence can simultaneously extract the fault reason and the characteristics of harmonic waves, amplitude values, slopes and the like of the distance volt-ampere characteristic curve image, a characteristic diagram is formed through the add layer of the model, the characteristics influencing fault reason identification are automatically extracted from the characteristic diagram through the full connection layer (fc _ 3), and the characteristics influencing fault distance identification are automatically extracted from the characteristic diagram through the full connection layer (fc _ 2). Preferably, the slope of the current-voltage characteristic curve related to the current and voltage of the transmission line reflects the fault distance. Compared with the prior art, in the test of the fault reason and the fault distance, the fault reason identification model and the fault distance identification model do not need to be independently arranged, the extraction depending on signal analysis and characteristic values is not needed, and the identification of the fault reason and the fault distance can be simultaneously completed based on the extraction of the characteristics of the volt-ampere characteristic curve image.
The test result is shown in fig. 7, and the result shows that the accuracy can reach 100% no matter whether the fault type is identified or the fault reason is identified, and the average error of the fault location is 0.285%.
By means of the application of the image recognition algorithm in the power system, the invention can realize the simultaneous learning of complete voltage and current signals by using the volt-ampere characteristic curve image recognition algorithm without depending on the mining of small differences of different faults by using a signal analysis method which is complicated in calculation and involves prior experience.
The invention constructs a label-free sample re-labeling method based on semi-supervised learning by combining class imbalance of transmission line faults, trains a deep learning neural network model, constructs model input by a volt-ampere characteristic curve, outputs fault classification labels and fault distances, finally forms a transmission line fault phase identification, reason and distance identifier, and applies the identifier to two ends of a transmission line, namely a leading-out terminal of a transformer substation, and is used for identifying and positioning the transmission line faults.
The second embodiment of the invention also discloses a transmission line fault identification and fault location device, which comprises:
the first acquisition unit is used for acquiring line data of the power transmission line;
the volt-ampere characteristic curve construction unit is used for constructing a volt-ampere characteristic curve of the power transmission line based on the acquired line data of the power transmission line;
the normalization processing unit is used for normalizing the volt-ampere characteristic curve constructed based on the line data of the power transmission line into a single total volt-ampere characteristic curve image which is used as a fault type identification volt-ampere characteristic curve graph;
the fault type identification unit is used for identifying the fault type identification volt-ampere characteristic curve according to a pre-constructed fault type identification model and outputting the identified fault type;
the image construction unit is used for obtaining a fault phase corresponding to the fault type of the power transmission line based on the identified fault type, and then constructing a single fault phase volt-ampere characteristic curve image corresponding to the fault phase as a fault reason and distance identification volt-ampere characteristic curve graph;
and the fault reason and distance identification unit is used for identifying and processing the fault reason and distance identification volt-ampere characteristic curve graph according to a pre-constructed fault reason and distance identification model so as to output the identified fault reason and fault distance.
Further, in the apparatus disclosed in the second aspect of the present invention, the apparatus further includes:
the model building unit is used for respectively building a fault type original identification model and a fault reason and distance original identification model before acquiring line data of the power transmission line;
the second acquisition unit is used for acquiring line data of a plurality of fault transmission lines for fault samples to form a fault sample data set;
the system comprises a sample set classification unit, a fault analysis unit and a fault analysis unit, wherein the sample set classification unit is used for acquiring a line fault type sample set and a line fault reason and distance sample set based on the fault sample data set;
the first training unit is used for inputting the obtained line fault sample set into the fault type original identification model to carry out transfer learning training until convergence, so as to obtain the fault type identification model; and the number of the first and second groups,
and the second training unit is used for inputting the obtained line fault reason and distance sample set into the fault reason and distance original identification model to perform transfer learning training until convergence, so as to obtain the fault reason and distance identification model.
A third embodiment of the present invention discloses an electronic device, which includes a memory and a processor, wherein the memory is used for storing a computer program, and the processor runs the computer program to make the electronic device execute the method for identifying and ranging the power transmission line fault described in the foregoing first aspect.
A fourth embodiment of the present invention 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 and ranging a power transmission line fault described in the first aspect of the present invention is performed.
In the several embodiments provided in the present application, it should be understood that the disclosed apparatus and method may be implemented in other manners. The apparatus embodiments described above are merely illustrative, and for example, the flowchart and block diagrams in the figures 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, functional modules in the embodiments of the present application may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent 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 or portions thereof that substantially contribute to the prior art may be embodied in the form of a software product stored in a storage medium and including instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to 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), a magnetic disk or an optical disk, and other various media capable of storing program codes.
The above description is only an example of the present application and is not intended to limit the scope of the present application, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present application shall be included in the protection scope of the present application. It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus, once an item is defined in one figure, it need not be further defined or explained in subsequent figures.
The above description is only for the 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 conceive of the changes or substitutions within the technical scope of the present application, and shall 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, herein, relational terms such as first and second, and the like may be 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. Also, 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 phrases "comprising a," "8230," "8230," or "comprising" does not exclude the presence of additional like elements in a process, method, article, or apparatus that comprises the element.
Finally, the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting, although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications or equivalent substitutions may be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all of them should be covered in the claims of the present invention.
Claims (10)
1. A method for identifying and ranging faults of a power transmission line is characterized by comprising the following steps: the method comprises the following steps:
acquiring line data of the power transmission line;
establishing a volt-ampere characteristic curve of the power transmission line based on the acquired line data of the power transmission line;
normalizing a volt-ampere characteristic curve constructed based on line data of a power transmission line into a single total volt-ampere characteristic curve image as a fault type identification volt-ampere characteristic curve graph;
identifying the fault type identification volt-ampere characteristic curve according to a fault type identification model which is constructed in advance, and outputting the identified fault type;
acquiring a fault phase corresponding to the fault type of the power transmission line based on the identified fault type, and then constructing a single fault phase volt-ampere characteristic curve image corresponding to the fault phase as a fault reason and distance identification volt-ampere characteristic curve;
and identifying the fault reason and distance identification volt-ampere characteristic curve according to a pre-constructed fault reason and distance identification model so as to output the identified fault reason and fault distance.
2. The method of claim 1, wherein the fault identification and ranging is performed by a mobile station,
the method for constructing the single fault phase volt-ampere characteristic curve image corresponding to the fault phase comprises the following steps: and processing the single total volt-ampere characteristic curve image based on the obtained fault phase corresponding to the fault type, and obtaining a single fault phase volt-ampere characteristic curve image corresponding to the fault phase.
3. The method of claim 1, wherein the fault identification and ranging is performed by a mobile station,
the line data comprises wave recording data of secondary side three-phase voltage and current signals at two ends of the power transmission line and wave recording data of zero sequence voltage and current signals;
the establishing of the volt-ampere characteristic curve of the power transmission line based on the acquired line data of the power transmission line comprises the following steps: and constructing a three-phase volt-ampere characteristic curve and a zero sequence volt-ampere characteristic curve at the head end side of the power transmission line, and constructing a three-phase volt-ampere characteristic curve and a zero sequence volt-ampere characteristic curve at the tail end side of the power transmission line.
4. The method of claim 1, wherein the fault identification and ranging is performed by a mobile station,
before acquiring line data of the power transmission line, the method further comprises:
constructing a fault type original identification model and a fault reason and distance original identification model;
acquiring line data of a plurality of power transmission lines for generating fault samples to form a fault sample data set;
acquiring a line fault type sample set and a line fault reason and distance sample set based on the fault sample data set;
inputting the obtained line fault sample set into the fault type original recognition model to perform transfer learning training until convergence, so as to obtain the fault type recognition model; and the number of the first and second groups,
and inputting the obtained line fault reason and distance sample set into the fault reason and distance original identification model to perform transfer learning training until convergence.
5. The transmission line fault identification and fault location method of claim 4,
the acquiring of the line fault type sample set includes:
establishing a volt-ampere characteristic curve of a corresponding fault power transmission line based on the acquired fault sample data set;
normalizing the built volt-ampere characteristic curve of the corresponding fault power transmission line into a single total volt-ampere characteristic curve image sample as a fault type identification volt-ampere characteristic curve sample;
acquiring a plurality of fault type identification volt-ampere characteristic curve chart samples to form a line fault type sample set;
the fault sample data set comprises a fault type mark, and the obtaining of the line fault reason and the distance sample set comprises the following steps:
processing the single total volt-ampere characteristic curve image sample based on a fault type mark in a fault sample data set, and obtaining a single fault phase volt-ampere characteristic curve image sample corresponding to a fault phase;
and acquiring a plurality of single fault phase voltage-current characteristic curve image samples to form a line fault reason and distance sample set.
6. The method for power transmission line fault identification and fault location according to claim 4,
the line data of the fault power transmission line comprises: recording data of secondary side three-phase voltage and current signals, recording data of secondary side zero sequence voltage and current signals, fault type marks, fault reason marks and fault distance marks at two ends of the power transmission line;
in the step of obtaining line data of a plurality of faulty transmission lines for generating faulty samples to form a faulty sample data set, the method includes:
taking the proportion of each fault type to the total fault type and the proportion of each fault reason to the total fault reason as classification weights;
in the step of obtaining line data of a plurality of faulty transmission lines for the faulty sample to form a faulty sample data set, the method further includes: a semi-supervised model is established to label unlabeled fault samples.
7. The method according to claim 4, wherein the original identification model of the fault type is a single-output convolutional neural network architecture, and the original identification model of the fault cause and distance is a multi-output convolutional neural network architecture.
8. The utility model provides a transmission line trouble is discerned and trouble range unit, its characterized in that, transmission line trouble is discerned and trouble range unit includes:
the first acquisition unit is used for acquiring line data of the power transmission line;
the volt-ampere characteristic curve construction unit is used for constructing a volt-ampere characteristic curve of the power transmission line based on the acquired line data of the power transmission line;
the normalization processing unit is used for normalizing the volt-ampere characteristic curve constructed based on the line data of the power transmission line into a single total volt-ampere characteristic curve image which is used as a fault type identification volt-ampere characteristic curve graph;
the fault type identification unit is used for identifying the fault type identification volt-ampere characteristic curve according to a pre-constructed fault type identification model and outputting the identified fault type;
the image construction unit is used for obtaining fault phases corresponding to the fault types of the power transmission line based on the identified fault types, and then constructing a single fault phase volt-ampere characteristic curve image corresponding to the fault phase as a fault reason and distance identification volt-ampere characteristic curve chart;
and the fault reason and distance identification unit is used for identifying and processing the fault reason and distance identification volt-ampere characteristic curve graph according to a pre-constructed fault reason and distance identification model so as to output the identified fault reason and fault distance.
9. The transmission line fault identification and fault location device of claim 8, further comprising:
the model building unit is used for respectively building a fault type original identification model and a fault reason and distance original identification model before acquiring line data of the power transmission line;
the second acquisition unit is used for acquiring line data of a plurality of fault transmission lines used for fault samples to form a fault sample data set;
the system comprises a sample set classification unit, a fault analysis unit and a fault analysis unit, wherein the sample set classification unit is used for acquiring a line fault type sample set, a line fault reason and a distance sample set based on a fault sample data set;
the first training unit is used for inputting the obtained line fault sample set into the fault type original recognition model to carry out transfer learning training until convergence, so as to obtain the fault type recognition model; and the number of the first and second groups,
and the second training unit is used for inputting the obtained line fault reason and distance sample set into the original fault reason and distance identification model to perform transfer learning training until convergence, so as to obtain the fault reason and distance identification model.
10. An electronic device, comprising a memory for storing a computer program and a processor for executing the computer program to cause the electronic device to perform the method of transmission line fault identification and fault location according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310114821.7A CN115932484B (en) | 2023-02-15 | 2023-02-15 | Power transmission line fault identification and fault location method and device and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310114821.7A CN115932484B (en) | 2023-02-15 | 2023-02-15 | Power transmission line fault identification and fault location method and device and electronic equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115932484A true CN115932484A (en) | 2023-04-07 |
| CN115932484B CN115932484B (en) | 2023-07-18 |
Family
ID=86697899
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310114821.7A Active CN115932484B (en) | 2023-02-15 | 2023-02-15 | Power transmission line fault identification and fault location method and device and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115932484B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116094179A (en) * | 2023-04-11 | 2023-05-09 | 国网安徽省电力有限公司合肥供电公司 | AC/DC flexible distribution network medium-voltage line fault analysis processing system |
| 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 (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106405339A (en) * | 2016-11-11 | 2017-02-15 | 中国南方电网有限责任公司 | Power transmission line fault reason identification method based on high and low frequency wavelet feature association |
| CN108344922A (en) * | 2017-12-30 | 2018-07-31 | 国网重庆市电力公司万州供电分公司 | A kind of transmission line of electricity direct lightning strike fault recognition method based on similar differentiation and S-transformation |
| CN108647479A (en) * | 2018-07-03 | 2018-10-12 | 广东电网有限责任公司 | A kind of surge arrester failure transient-wave diagnostic method and device |
| CN108959732A (en) * | 2018-06-15 | 2018-12-07 | 西安科技大学 | A kind of transmission line malfunction kind identification method based on convolutional neural networks |
| CN109270407A (en) * | 2018-11-16 | 2019-01-25 | 国网山东省电力公司电力科学研究院 | Extra high voltage direct current transmission line failure cause discrimination method based on Multi-source Information Fusion |
| CN109324266A (en) * | 2018-11-21 | 2019-02-12 | 国网电力科学研究院武汉南瑞有限责任公司 | A kind of distribution single-phase-to-earth fault analysis method based on deep learning |
| CN111046581A (en) * | 2019-12-27 | 2020-04-21 | 国网江苏省电力有限公司电力科学研究院 | Power transmission line fault type identification method and system |
| CN111881971A (en) * | 2020-07-24 | 2020-11-03 | 成都理工大学 | Power transmission line fault type identification method based on deep learning LSTM model |
| CN112147462A (en) * | 2020-09-16 | 2020-12-29 | 国网江西省电力有限公司电力科学研究院 | Power transmission line fault identification method based on deep learning |
| CN112255499A (en) * | 2020-10-10 | 2021-01-22 | 重庆大学 | Phase current amplitude based power distribution network disconnection fault positioning and identifying method and system |
| US20210048487A1 (en) * | 2019-08-12 | 2021-02-18 | Wuhan University | Power transformer winding fault positioning method based on deep convolutional neural network integrated with visual identification |
| CN113011084A (en) * | 2021-02-26 | 2021-06-22 | 华北电力大学 | Intelligent identification method for transmission line fault reasons based on correlation vector machine |
| US20210293873A1 (en) * | 2020-03-18 | 2021-09-23 | Mitsubishi Electric Research Laboratories, Inc. | Transient based Fault Location Method for Ungrounded Power Distribution Systems |
| CN114034980A (en) * | 2021-11-12 | 2022-02-11 | 陕西省地方电力(集团)有限公司渭南供电分公司 | Power distribution line fault detection method based on particle swarm optimization BP neural network |
| CN114047705A (en) * | 2021-12-07 | 2022-02-15 | 国网河南省电力公司电力科学研究院 | Single-phase earth fault detection constant value self-adaptive setting method and system |
| CN114062832A (en) * | 2021-08-31 | 2022-02-18 | 广东电网有限责任公司 | Method and system for identifying short-circuit fault type of power distribution network |
| CN114117921A (en) * | 2021-11-29 | 2022-03-01 | 青海大学 | Intelligent diagnosis method for faults of photovoltaic array |
| CN114355099A (en) * | 2021-12-07 | 2022-04-15 | 国网河南省电力公司电力科学研究院 | Power distribution network single-phase earth fault detection method based on field wave recording data analysis |
| 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 |
| US20220196720A1 (en) * | 2020-12-18 | 2022-06-23 | Wuhan University | Single-ended fault positioning method and system for high-voltage direct-current transmission line of hybrid network |
| CN114755529A (en) * | 2022-04-06 | 2022-07-15 | 重庆大学 | Multi-feature fusion single-phase earth fault type identification method based on deep learning |
| US20220268827A1 (en) * | 2021-02-24 | 2022-08-25 | Mitsubishi Electric Research Laboratories, Inc. | Distribution Fault Location Using Graph Neural Network with both Node and Link Attributes |
| CN115469184A (en) * | 2022-09-05 | 2022-12-13 | 广东电网有限责任公司广州供电局 | New energy transmission line fault identification method based on convolutional network |
-
2023
- 2023-02-15 CN CN202310114821.7A patent/CN115932484B/en active Active
Patent Citations (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106405339A (en) * | 2016-11-11 | 2017-02-15 | 中国南方电网有限责任公司 | Power transmission line fault reason identification method based on high and low frequency wavelet feature association |
| CN108344922A (en) * | 2017-12-30 | 2018-07-31 | 国网重庆市电力公司万州供电分公司 | A kind of transmission line of electricity direct lightning strike fault recognition method based on similar differentiation and S-transformation |
| CN108959732A (en) * | 2018-06-15 | 2018-12-07 | 西安科技大学 | A kind of transmission line malfunction kind identification method based on convolutional neural networks |
| CN108647479A (en) * | 2018-07-03 | 2018-10-12 | 广东电网有限责任公司 | A kind of surge arrester failure transient-wave diagnostic method and device |
| CN109270407A (en) * | 2018-11-16 | 2019-01-25 | 国网山东省电力公司电力科学研究院 | Extra high voltage direct current transmission line failure cause discrimination method based on Multi-source Information Fusion |
| CN109324266A (en) * | 2018-11-21 | 2019-02-12 | 国网电力科学研究院武汉南瑞有限责任公司 | A kind of distribution single-phase-to-earth fault analysis method based on deep learning |
| US20210048487A1 (en) * | 2019-08-12 | 2021-02-18 | Wuhan University | Power transformer winding fault positioning method based on deep convolutional neural network integrated with visual identification |
| CN111046581A (en) * | 2019-12-27 | 2020-04-21 | 国网江苏省电力有限公司电力科学研究院 | Power transmission line fault type identification method and system |
| US20210293873A1 (en) * | 2020-03-18 | 2021-09-23 | Mitsubishi Electric Research Laboratories, Inc. | Transient based Fault Location Method for Ungrounded Power Distribution Systems |
| CN111881971A (en) * | 2020-07-24 | 2020-11-03 | 成都理工大学 | Power transmission line fault type identification method based on deep learning LSTM model |
| CN112147462A (en) * | 2020-09-16 | 2020-12-29 | 国网江西省电力有限公司电力科学研究院 | Power transmission line fault identification method based on deep learning |
| CN112255499A (en) * | 2020-10-10 | 2021-01-22 | 重庆大学 | Phase current amplitude based power distribution network disconnection fault positioning and identifying method and system |
| US20220196720A1 (en) * | 2020-12-18 | 2022-06-23 | Wuhan University | Single-ended fault positioning method and system for high-voltage direct-current transmission line of hybrid network |
| US20220268827A1 (en) * | 2021-02-24 | 2022-08-25 | Mitsubishi Electric Research Laboratories, Inc. | Distribution Fault Location Using Graph Neural Network with both Node and Link Attributes |
| CN113011084A (en) * | 2021-02-26 | 2021-06-22 | 华北电力大学 | Intelligent identification method for transmission line fault reasons based on correlation vector machine |
| CN114062832A (en) * | 2021-08-31 | 2022-02-18 | 广东电网有限责任公司 | Method and system for identifying short-circuit fault type of power distribution network |
| CN114034980A (en) * | 2021-11-12 | 2022-02-11 | 陕西省地方电力(集团)有限公司渭南供电分公司 | Power distribution line fault detection method based on particle swarm optimization BP neural network |
| CN114117921A (en) * | 2021-11-29 | 2022-03-01 | 青海大学 | Intelligent diagnosis method for faults of photovoltaic array |
| CN114355099A (en) * | 2021-12-07 | 2022-04-15 | 国网河南省电力公司电力科学研究院 | Power distribution network single-phase earth fault detection method based on field wave recording data analysis |
| CN114047705A (en) * | 2021-12-07 | 2022-02-15 | 国网河南省电力公司电力科学研究院 | Single-phase earth fault detection constant value self-adaptive setting method and system |
| 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 |
| CN114755529A (en) * | 2022-04-06 | 2022-07-15 | 重庆大学 | Multi-feature fusion single-phase earth fault type identification method based on deep learning |
| CN115469184A (en) * | 2022-09-05 | 2022-12-13 | 广东电网有限责任公司广州供电局 | New energy transmission line fault identification method based on convolutional network |
Non-Patent Citations (9)
| Title |
|---|
| AHMAD ABDULLAH: "Ultrafast Transmission Line Fault Detection Using a DWT-Based ANN", 《IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS》, pages 1182 - 1192 * |
| ALEX KRIZHEVSKY，ILYA SUTSKEVER，GEOFFREY E. HINTON: "ImageNet Classification with Deep Convolutional Neural Networks", 《RESEARCH HIGHLIGHTS》, pages 84 - 89 * |
| MOSLEM SALEHI,FARHAD NAMDARI.: "Fault classification and faulted phase selection for transmission line using morphological edge detection filter", 《IET JOURNALS》, pages 1595 - 1604 * |
| 余锐;肖超;陈愚;欧阳金鑫;熊俊;熊小伏: "柔性直流馈入下交流输电线路单端故障测距分析", 《重庆大学学报》, pages 48 - 55 * |
| 徐舒玮;邱才明;张东霞;贺兴;储磊;杨浩森: "基于深度学习的输电线路故障类型辨识", 《中国电机工程学报》, pages 1 - 11 * |
| 李昊; 王建; 熊小伏; 张波; 陈红州: "基于双目图像测距的输电线路净距计算与安全告警方法", 《广东电力》, pages 11 - 19 * |
| 杨毅; 范栋琛; 殷浩然; 韩佶; 苗世洪: "基于深度-迁移学习的输电线路故障选相模型及其可迁移性研究", 《电力自动化设备》, pages 1 - 8 * |
| 王建; 姚江宁; 刘泽青; 欧阳金鑫; 熊小伏: "恶劣天气下配电网故障统计分析及其概率分布拟合", 《电力系统保护与控制》, pages 144 - 152 * |
| 王建;吴昊;张博;南东亮;欧阳金鑫;熊小伏: "不平衡样本下基于迁移学习-AlexNet 的输电线路故障辨识方法", 《电力系统自动化》, pages 182 - 189 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116094179A (en) * | 2023-04-11 | 2023-05-09 | 国网安徽省电力有限公司合肥供电公司 | AC/DC flexible distribution network medium-voltage line fault analysis processing system |
| 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 |
|---|---|
| CN115932484B (en) | 2023-07-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115932484B (en) | Power transmission line fault identification and fault location method and device and electronic equipment | |
| CN110082640B (en) | Distribution network single-phase earth fault identification method based on long-time memory network | |
| Farshad et al. | A novel fault-location method for HVDC transmission lines based on similarity measure of voltage signals | |
| CN112147462A (en) | Power transmission line fault identification method based on deep learning | |
| Ferreira et al. | Probabilistic transmission line fault diagnosis using autonomous neural models | |
| CN112787591B (en) | Photovoltaic array fault diagnosis method based on fine-tuning dense connection convolutional neural network | |
| 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 | |
| CN116008731B (en) | Power distribution network high-resistance fault identification method and device and electronic equipment | |
| CN113447766A (en) | Method, device, equipment and storage medium for detecting high-resistance ground fault | |
| CN115545479A (en) | Method and device for determining important nodes or important lines of power distribution network | |
| Shi et al. | Diagnosis of the single phase‐to‐ground fault in distribution network based on feature extraction and transformation from the waveforms | |
| CN111999591B (en) | Method for identifying abnormal state of primary equipment of power distribution network | |
| CN113610119B (en) | Method for identifying power transmission line development faults based on convolutional neural network | |
| Mehinović et al. | Application of artificial intelligence methods for determination of transients in the power system | |
| Wu et al. | Transmission line fault cause identification method based on transient waveform image and MCNN-LSTM | |
| Asbery et al. | Electric transmission system fault identification using modular artificial neural networks for single transmission lines | |
| Mahela et al. | A protection scheme for distribution utility grid with wind energy penetration | |
| Muzzammel et al. | Low impedance fault identification and classification based on Boltzmann machine learning for HVDC transmission systems | |
| CN114595746A (en) | Method, device, equipment and storage medium for classifying power distribution network fault reasons | |
| Cui et al. | High impedance fault detection method based on Data divergence in the distribution network | |
| França et al. | A machine learning-based approach for comprehensive fault diagnosis in transmission lines | |
| Venkata et al. | Data Mining and SVM Based Fault Diagnostic Analysis in Modern Power System Using Time and Frequency Series Parameters Calculated From Full-Cycle Moving Window | |
| CN117250439B (en) | Three-layer type studying and judging analysis system for multi-source ground fault | |
| Khoudry et al. | Empirical mode decomposition and cascade feed-forward artificial neural network based intelligent fault classifier |
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 |