Fault identification method, device, electronic equipment and storage medium for transmission line

Abstract

The embodiment of the present application is applicable to the field of power transmission network technology, and provides a method, device, electronic device and storage medium for identifying faults in power transmission lines. The method includes: obtaining fault recording data and meteorological data in multiple consecutive cycles when a fault occurs in a power transmission line; performing feature extraction on the fault recording data and meteorological data corresponding to the multiple cycles, respectively, to obtain the first fault features and first meteorological features corresponding to the multiple cycles; predicting the fault results of the power transmission line based on the multiple first fault features and meteorological features. The above method can accurately determine the cause of the power transmission line fault, so as to provide guidance to the line patrol personnel when troubleshooting the power transmission line.

Description

Technical Field

[0001] The present application belongs to the technical field of power transmission networks, and in particular, relates to a method, device, electronic equipment and storage medium for identifying faults in power transmission lines. Background Art

[0002] Due to the uneven distribution of energy and load centers in various places, it is usually necessary to use transmission lines to achieve long-distance transmission of energy and load. Among them, since transmission lines are exposed to the wild for a long time, they are easily affected by severe weather such as lightning, ice and snow, strong winds, and often fail. In addition, when eliminating faults caused by different fault results, the fault elimination solutions are usually different.

[0003] At present, after a transmission line fault occurs, it is usually necessary for line patrol personnel to gradually eliminate the transmission line to determine the possible causes of the fault, and then formulate a reasonable plan to eliminate the fault.

[0004] However, since it is impossible to determine the possible cause of the fault, it is impossible to provide guidance to the line patrol personnel when they are troubleshooting the transmission line. As a result, it is easy for the line patrol personnel to spend a lot of time to determine the cause, which has a great impact on the social economy and people's lives. Summary of the invention

[0005] The embodiments of the present application provide a method, device, electronic device and storage medium for identifying a fault in a power transmission line, which can solve the problem that the cause of the power transmission line fault cannot be determined in advance.

[0006] In a first aspect, an embodiment of the present application provides a method for identifying a fault of a power transmission line, the method comprising:

[0007] Obtain fault recording data and meteorological data for multiple consecutive cycles when a transmission line fault occurs;

[0008] Extracting features from the fault recording data and meteorological data corresponding to the multiple periods respectively, and obtaining first fault features and first meteorological features corresponding to the multiple periods respectively;

[0009] The failure result of the transmission line is predicted based on the plurality of first fault characteristics and the meteorological characteristics.

[0010] In a second aspect, an embodiment of the present application provides a transmission line fault identification device, the device comprising:

[0011] An acquisition module is used to acquire fault recording data and meteorological data in multiple consecutive cycles when a transmission line fault occurs;

[0012] A feature extraction module is used to extract features from the fault recording data and meteorological data corresponding to the multiple periods, respectively, to obtain first fault features and first meteorological features corresponding to the multiple periods respectively;

[0013] The prediction module is used to predict the failure result of the transmission line according to a plurality of first fault characteristics and meteorological characteristics.

[0014] In a third aspect, an embodiment of the present application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method of the first aspect described above when executing the computer program.

[0015] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the method of the first aspect described above is implemented.

[0016] In a fifth aspect, an embodiment of the present application provides a computer program product, which, when executed on an electronic device, enables the electronic device to execute the method of the first aspect.

[0017] Compared with the prior art, the embodiment of the present application has the following beneficial effects: the electronic device can obtain the fault recording data and meteorological data in multiple continuous cycles when the transmission line fails, and extract the features of the fault recording data and meteorological data corresponding to the multiple cycles respectively, and obtain the first fault features and first meteorological features corresponding to the multiple cycles respectively. Finally, according to the multiple first fault features and meteorological features, the fault result of the transmission line is predicted. Based on this, the fault recording data and the meteorological data that may cause the transmission line failure are combined to predict the fault result of the transmission line through multi-dimensional data features, which can improve the prediction accuracy. Moreover, the fault recording data and meteorological data are both time series data in multiple continuous cycles. Therefore, it can be considered that the corresponding multiple first fault features and first meteorological features can describe the time series feature transformation law of the fault recording data when the fault occurs. Furthermore, on the basis that the faults caused by different external factors may cause the transmission line to produce different fault recording data, the fault result is predicted based on the time series feature transformation law, which can further improve the prediction accuracy of the fault result. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0019] Figure 1 This is a flow chart of a method for identifying a fault in a power transmission line provided in one embodiment of the present application;

[0020] Figure 2 It is a schematic diagram of the relationship between the zero-sequence DC component and the fault result in a method for identifying a fault of a power transmission line provided in one embodiment of the present application;

[0021] Figure 3 This is a schematic diagram of the relationship between the third harmonic of the zero-sequence current and the fault result in a method for identifying a fault of a power transmission line provided in an embodiment of the present application;

[0022] Figure 4 It is a schematic diagram of the relationship between the third harmonic of the zero-sequence voltage and the fault result in a method for identifying a fault of a power transmission line provided in an embodiment of the present application;

[0023] Figure 5 It is a schematic diagram of the relationship between the simulated maximum distribution of wavelet mode energy and the fault result in a method for identifying a fault of a power transmission line provided in an embodiment of the present application;

[0024] Figure 6 This is a schematic diagram of the relationship between the simulated maximum distribution of wavelet modulus extreme values ​​and the fault result in a method for identifying a power transmission line fault provided in one embodiment of the present application;

[0025] Figure 7 It is a schematic diagram of the relationship between the simulated maximum distribution of the wavelet modulus average value and the fault result in a method for identifying a power transmission line fault provided in one embodiment of the present application;

[0026] Figure 8 It is a schematic diagram of the relationship between the simulated maximum distribution of the wavelet mode standard deviation and the fault result in a method for identifying a power transmission line fault provided in one embodiment of the present application;

[0027] Fig. 9 This is a schematic diagram of an implementation method for predicting fault results in a method for identifying a fault in a power transmission line provided in an embodiment of the present application;

[0028] Fig.10 This is a flow chart of a method for identifying a fault in a power transmission line provided in one embodiment of the present application;

[0029] Fig.11 It is a structural schematic diagram of a fault identification device for a power transmission line provided in one embodiment of the present application;

[0030] Fig.12 It is a structural schematic diagram of an electronic device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0031] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

[0032] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or combinations thereof.

[0033] It should be noted that the information collection process (such as face image collection process, fingerprint information collection process, etc.) / feature extraction process involved in this application is performed with the user's knowledge and permission, that is, the information collection process / feature extraction process complies with the requirements of laws and regulations and does not constitute an act that harms the public interest.

[0034] In addition, in the description of the present application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0035] Due to the uneven distribution of energy and load centers in various places, it is usually necessary to use transmission lines to achieve long-distance transmission of energy and load. Among them, since transmission lines are exposed to the wild for a long time, they are easily affected by severe weather such as lightning, ice and snow, strong winds, and often fail. In addition, when eliminating faults caused by different fault results, the fault elimination solutions are usually different.

[0036] Specifically, the causes of power transmission line failures are usually external force damage (e.g., crane spraying, tree obstacles), lightning strikes, wildfires, dirt, ice, bird strikes, etc. After the failure occurs, since power transmission line failures often occur in suburban areas, it may take several hours from tripping to restoring normal power supply of the line. In turn, it will have a great impact on the social economy and people's lives.

[0037] At present, after a transmission line fault occurs, it is usually necessary for line patrol personnel to gradually eliminate the transmission line to determine the possible causes of the fault, and then formulate a reasonable plan to eliminate the fault.

[0038] However, since it is impossible to determine the possible cause of the fault, it is impossible to provide guidance to the line patrol personnel when they are troubleshooting the transmission line. As a result, it is easy for the line patrol personnel to spend a lot of time on troubleshooting, which has a great impact on the social economy and people's lives.

[0039] Based on this, in order to determine the cause of the fault and assist line patrol personnel in troubleshooting, an embodiment of the present application provides a fault identification method for a power transmission line. The method can be applied to electronic devices such as tablet computers, laptop computers, and ultra-mobile personal computers (UMPC). The embodiment of the present application does not impose any restrictions on the specific type of electronic device.

[0040] See also Figure 1 , Figure 1 A flowchart of a method for identifying a fault in a power transmission line provided in an embodiment of the present application is shown. The method comprises the following steps:

[0041] S101. Acquire fault recording data and meteorological data in multiple consecutive cycles when a transmission line fault occurs.

[0042] In one embodiment, the possible faults of the power transmission line can refer to the fault results described above, and no further explanation is given.

[0043] In one embodiment, the above cycle can be considered as a process in which the voltage signal or current signal changes periodically over time. The cycle can be a preset fixed cycle. For example, taking alternating current as an example, the cycle of alternating current is 50 Hz, that is, the alternating current wave repeats 50 times per second.

[0044] The number of cycles can be set according to actual conditions and is not limited thereto. For example, it can be 5 times. Furthermore, the fault recording data and meteorological data corresponding to the multiple cycles must have temporal continuity.

[0045] It can be understood that the fault recording data with time continuity can usually more accurately characterize the changes in fault data when a transmission line fault occurs.

[0046] In this embodiment, in order to better characterize the change of fault data when a transmission line fails, the multiple cycles include at least one cycle before the failure and at least one cycle after the failure. Furthermore, it can be considered that the fault recording data and meteorological data corresponding to the multiple cycles can characterize the data change from normal to failure of the transmission line in detail and accurately.

[0047] In one embodiment, the above-mentioned fault recording data includes but is not limited to one or more data such as phase voltage, phase current, zero-sequence voltage, zero-sequence current, transition resistance, etc., and no limitation is made to this.

[0048] Among them, the electronic device can obtain the above-mentioned fault recording data through a fault recording device. Specifically, the fault recording device is an automated device for monitoring the operation status of the power system. When a fault occurs in the power system (including the transmission line), the fault recording device can record the fault to directly record various fault data reflecting the fault of the power system and generate a recording file. Afterwards, the electronic device can perform waveform analysis on the recording file to obtain the corresponding fault recording data.

[0049] Among them, waveform analysis of the recording file can usually obtain one or more information such as the type of fault, phase difference and its evolution process, the action of relay protection, the working condition of the circuit breaker during the fault and the fault location.

[0050] Exemplarily, the information that can be obtained by waveform analysis of the recording file includes but is not limited to: the faulty electrical component and fault type, protection action time and fault clearing time, fault current and fault voltage, reclosing time and whether reclosing is successful, detailed protection action status, impedance trajectory, phasor and harmonic analysis and other information.

[0051] In one embodiment, the amplitude and phase changes of the fault phase voltage and phase current are mostly affected by the position of the line fault point and the size of the transition resistance, and the waveform changes of the fault phase voltage and phase current are essentially determined by the properties of the transition resistance.

[0052] In addition, the fault reclosing characteristics have a certain correspondence with the cause of the fault. For example, for instantaneous faults caused by lightning strikes, pollution flashovers, etc., reclosing is easy to succeed; for faults caused by forest fires, cranes hitting wires, etc., reclosing is not easy to succeed.

[0053] In addition, the high-frequency harmonic content and DC component of the fault phase current and zero-sequence current are also related to the fault results. Figure 2 , Figure 3 and Figure 4 , Figure 2 It is a schematic diagram of the relationship between the zero-sequence DC component and the fault result in a method for identifying a fault of a power transmission line provided in one embodiment of the present application; Figure 3 is a schematic diagram of the relationship between the third harmonic of zero-sequence current and the fault result in a method for identifying a fault of a power transmission line provided in an embodiment of the present application; and, Figure 4 It is a schematic diagram of the relationship between the third harmonic of the zero-sequence voltage and the fault result in a fault identification method for a power transmission line provided in an embodiment of the present application.

[0054] Among them, the above Figure 2 , Figure 3 and Figure 4The fault recording data of 200 transmission lines were extracted and Fourier transformed to obtain the zero-sequence DC component, zero-sequence current third harmonic, and zero-sequence voltage third harmonic data distribution. Based on this, it can be determined according to the above multiple figures:

[0055] The zero-sequence DC component corresponding to the fault results such as lightning strike, wind deflection, bird flash, foreign object (for example, foreign object wiring) is relatively high. In addition, the zero-sequence DC average value corresponding to the lightning strike fault result is the highest, and the zero-sequence DC average value corresponding to the icing fault result is relatively low;

[0056] The third harmonic content of zero-sequence current corresponding to fault results such as wildfire and foreign objects is relatively high, while the third harmonic content of zero-sequence current corresponding to other fault results is relatively low;

[0057] In addition, the third harmonic content of the zero-sequence voltage corresponding to fault results such as wind deviation and foreign objects is relatively high, while the third harmonic content of the zero-sequence voltage corresponding to fault results such as lightning strike, icing, and bird flash is relatively low.

[0058] In another embodiment, when a transmission line fails, the voltage signal or current signal in the fault recording data will be disturbed, causing the voltage or current to suddenly change to form a peak to form an extreme point. Based on this, in this embodiment, waveform analysis of the recording file also includes wavelet change analysis of the recording file.

[0059] Specifically, in the wavelet transform, the singular point obtained by the modulus maximum corresponds to the mutation point of the acquired transient signal. In the power transmission line fault detection technology, one of the important technologies is to determine the specific location of the fault point. In this embodiment, it can be considered that the singular point information of the transient signal represents the location of the fault point. Therefore, the voltage signal or current signal in the recording file is decomposed and reconstructed by wavelet, and the signal components distributed in each frequency band can be obtained. Then, the signal components obtained after 3-layer decomposition and reconstruction using the DB4 wavelet basis are compared to analyze the relationship between the signal components and the fault results.

[0060] Specifically, refer to Figure 5 , Figure 6 , Figure 7 and Figure 8 , Figure 5 It is a schematic diagram of the relationship between the simulated maximum distribution in the wavelet mode energy and the fault result in a method for identifying a fault of a power transmission line provided in an embodiment of the present application; Figure 6 This is a schematic diagram of the relationship between the simulated maximum distribution of wavelet modulus extreme values ​​and the fault result in a method for identifying a power transmission line fault provided in one embodiment of the present application; Figure 7 is a schematic diagram of the relationship between the simulated maximum distribution of the wavelet modulus average value and the fault result in a method for identifying a power transmission line fault provided in an embodiment of the present application; and, Figure 8It is a schematic diagram of the relationship between the simulated maximum distribution of the wavelet modulus standard deviation and the fault result in a transmission line fault identification method provided in one embodiment of the present application.

[0061] Among them, the above Figure 5 , Figure 6 , Figure 7 and Figure 8 In order to select a recording file of a cycle at the beginning of the transmission line fault tripping for analysis, a schematic diagram of the distribution of the modulus maximum value in the wavelet mode energy, wavelet mode extreme value, wavelet mode average value and wavelet mode standard deviation is obtained. Based on the above multiple figures, it can be determined that:

[0062] Among them, the above Figures 5 to 8 In the multiple figures, 1 represents lightning strike, 2 represents wind deviation, 3 represents bird flash, 4 represents ice cover, 5 represents tree obstacle, 6 represents mountain fire, 7 represents foreign body wiring, etc. The modulus maximum distribution of the wavelet signal components corresponding to the fault results such as lightning strike, foreign body wiring, and ice cover is determined from the above multiple figures. The modulus maximum distribution of the wavelet signal components corresponding to the fault results such as lightning strike, foreign body wiring, and ice cover is at a relatively low level; and the modulus maximum distribution of the wavelet signal components corresponding to the fault results such as wind deviation, bird flash, mountain fire, and tree obstacle (external force factor) is at a relatively high level.

[0063] To summarize the above examples, in order to accurately predict the fault result based on the fault recording data, in this embodiment, data related to the fault result such as transition resistance value, transition resistance volt-ampere characteristic, reclosing characteristic, fault current DC content, fault current harmonic content, high-frequency wavelet modulus maximum value, high-frequency wavelet modulus standard deviation, high-frequency wavelet modulus average value can be used as fault recording data.

[0064] As an example, a data set corresponding to some fault recording data and fault results may be shown in Table 1 below:

[0065] Table 1:

[0066]

[0067] In one embodiment, the above-mentioned meteorological data includes but is not limited to one or more data such as weather type, time period, season, etc. Exemplarily, the weather type can be divided into one or more data such as thunderstorm, heavy snow, sunny day, etc. Time periods can be divided into: early morning (5:00-9:00), daytime (10:00-16:00), evening (17:00-22:00), midnight (23:00-4:00), etc.; seasons can be divided into four categories: spring (March-May), summer (June-August), autumn (September-November), winter (December-February), etc.

[0068] It should be noted that transmission line failures are usually greatly affected by weather conditions. For example, wildfires, bird strikes and other failures often occur on sunny days; lightning failures usually occur when transmission lines are operating under thunderstorm conditions and are affected by lightning activities; wildfire failures may also be caused by careless human behavior, so failures often occur during daytime hours; and some failures may also be related to natural laws and have a certain seasonality. For example, lightning activities are more frequent in summer and autumn, and the possibility of lightning failures is higher.

[0069] Specifically, the correlation between each fault and meteorological data can be shown in Table 2 below:

[0070] Table 2:

[0071]

[0072] In summary, in order to accurately predict the failure result, in this embodiment, three factors such as weather type, time period, and season can be selected as meteorological data to participate in subsequent processing.

[0073] It should be noted that the duration of a cycle of voltage or current in a transmission line is relatively short, and meteorological data usually does not change significantly in a short period of time. Therefore, it can be considered that the meteorological data in multiple consecutive cycles are the same.

[0074] Based on this, in this embodiment, only one set of the above-mentioned meteorological data needs to be obtained, and there is no need to obtain a set of meteorological data in each cycle.

[0075] S102 , extracting features from the fault recording data and meteorological data corresponding to the multiple periods respectively, to obtain first fault features and first meteorological features corresponding to the multiple periods respectively.

[0076] In one embodiment, the method of extracting features from fault recording data and meteorological data includes but is not limited to: one hot encoding, label encoding, hash encoding, etc., one or more methods, which are not limited to this.

[0077] In this embodiment, in order to facilitate the extracted first fault feature and first meteorological feature to be more suitable for recognition by the neural network model, one-hot encoding can be used for feature extraction, wherein one-hot encoding can represent each category as a unique vector, making it easier for the model to learn the features of each category.

[0078] For example, for fault categories such as lightning strike, wildfire, and bird strike, one-hot encoding can represent them as features such as [1, 0, 0], [0, 1, 0], and [0, 0, 1], respectively. For example, [1, 0, 0] can represent that the feature meets the lightning strike fault, [0, 1, 0] can represent that the feature meets the wildfire fault, and [0, 0, 1] can represent that the feature meets the bird strike fault.

[0079] It can be understood that for each cycle of fault recording data, if the fault recording data is transition resistance volt-ampere characteristics, fault current DC content, fault current harmonic content, reclosing characteristics, transition resistance value, high-frequency wavelet modulus maximum value, high-frequency wavelet modulus standard deviation, high-frequency wavelet modulus average value and other data, then there will be 8 corresponding numerical features in one cycle. At this time, when the number of cycles is 5, there will be 5 groups of time-continuous numerical features, and the number of each group of numerical features is 8.

[0080] S103: predicting the failure result of the power transmission line according to the plurality of first fault characteristics and meteorological characteristics.

[0081] In one embodiment, the electronic device may input a plurality of first fault characteristics and meteorological characteristics into a preset neural network model for processing to predict the fault result of the power transmission line. The neural network model includes but is not limited to a deep neural network (DNN), a convolutional neural network (CNN), a recurrent neural network (RNN), and a long- and short-term memory network (LSTM), etc., and is not limited to this.

[0082] Among them, since the LSTM model has better generalization performance for multi-source data with time series characteristics, in order to improve the prediction accuracy of fault results, multiple first fault characteristics and meteorological characteristics can be input into the preset long short-term memory network model to obtain the fault results.

[0083] In one embodiment, the LSTM model can be pre-trained based on training data consisting of fault recording data and meteorological data in multiple cycles.

[0084] As an example, the long short-term memory network model includes a plurality of first sequential network modules connected in sequence, and the electronic device can Fig. 9 The steps S901-S904 shown predict the failure results. The details are as follows:

[0085] S901. For any first sequential network module, determine a first target fault feature and a first meteorological feature corresponding to a first period according to a first connection sequence of the current first sequential network module.

[0086] In one embodiment, the time sequence of the first cycle in the plurality of cycles is the same as the first connection sequence. The long short-term memory network model described above includes a plurality of first sequential network modules connected in sequence, each of which can perform one or more operations such as convolution and pooling on the input features, and then output the corresponding features to the next first sequential network module.

[0087] Therefore, it can be considered that in the long short-term memory network model, the first connection order of each first temporal network module for extracting features from data is fixed.

[0088] Based on the above description, it can be considered that the first target fault feature and the first meteorological feature input into the first first sequential network module are the first fault feature and the meteorological feature corresponding to the first cycle (first cycle). And, the first target fault feature and the first meteorological feature input into the second first sequential network module (in another embodiment, since the meteorological features of each cycle in a short period of time are the same, there is no need to input again) are the first fault feature and the meteorological feature corresponding to the second cycle. And so on, no detailed description is given for this.

[0089] S902: Obtain a second timing feature output by a previous first timing network module.

[0090] In one embodiment, based on the explanation of the first connection sequence in S901 above, it can be known that the second timing feature output by the previous first timing network module will also be used as the input of the current first timing network module.

[0091] S903: Output a third time series feature according to the second time series feature, the first target fault feature and the first meteorological feature.

[0092] The third timing feature is the second timing feature input to the next first timing network module. That is, when the next first timing network module is the first timing network module currently processing data, the third timing feature will be used as the input of the first timing network module. That is, as the second timing feature of the first timing network module.

[0093] It should be noted that the input of the current first time series network module is the second time series feature output by the previous first time series network module, as well as the first target fault feature and the first meteorological feature corresponding to the first cycle. Based on this, it can be considered that the third time series feature output after being processed by the current first time series network module can be used in the long short-term memory network model to further understand the time series relationship between the fault recording data and the meteorological data of two adjacent cycles, as well as the transformation relationship between the fault recording data. Classification based on the third time series feature that can describe the time series relationship between the fault recording data and the meteorological data, as well as the transformation relationship between the fault recording data, can improve the prediction accuracy of the fault result.

[0094] S904: predict the fault result based on the last second time sequence feature output by the last first time sequence network module.

[0095] In one embodiment, after the above steps S901-903 are processed, the last second time series feature output by the last first time series network module will be able to simultaneously describe the time series relationship between the fault recording data and the meteorological data of multiple cycles, as well as the fault recording data transformation relationship. Furthermore, the prediction of the fault result based on the last second time series feature can improve the prediction accuracy.

[0096] As an example, the long short-term memory network model may include multiple classifiers to output the prediction probability corresponding to each fault result distribution. Then, the result corresponding to the maximum value among the multiple prediction probabilities is determined as the above fault result.

[0097] In another embodiment, since the fault recording data and meteorological data in a continuous period usually have multicollinearity, the prediction space is unstable, resulting in weak generalization ability of the model. In addition, the multi-dimensional last second time series features are sparse, making it difficult for the classifier in the long short-term memory network model to find important data features. That is, the potential relationship between features may be ignored in the prediction process, reducing the prediction accuracy of the model.

[0098] Based on this, in order to improve the prediction accuracy of the model, the last second time series feature can be reduced in dimension to obtain the fourth time series feature. Then, the fault result is predicted based on the fourth time series feature.

[0099] The dimensionality reduction methods include, but are not limited to, multidimensional scaling, isometric feature mapping, principal component analysis, etc., which are not limited thereto.

[0100] In another embodiment, the long short-term memory model may further include a plurality of second sequential network modules connected in sequence, and the number of the second sequential network modules may be the same as the number of the first sequential network modules. In this case, the long short-term memory model may predict the failure result based on the plurality of first sequential network modules and the plurality of second sequential network modules.

[0101] Specifically, for any second sequential network module, the electronic device can determine the second target fault feature and the second meteorological feature corresponding to the second cycle according to the second connection sequence of the current second sequential network module; the time sequence of the second cycle in multiple cycles is the same as the second connection sequence. Then, the fifth sequential feature output by the previous second sequential network module is obtained; the sixth sequential feature is output according to the fifth sequential feature, the second target fault feature and the second meteorological feature; the sixth sequential feature is the fifth sequential feature input to the next second sequential network module. Finally, based on the last sixth sequential feature and the last second sequential feature output by the last second sequential network module, the fault result is predicted.

[0102] Among them, the way in which multiple second timing network modules process multiple first fault characteristics and meteorological characteristics to obtain the last sixth timing characteristic can be similar to the way in which multiple second timing network modules process multiple first fault characteristics and meteorological characteristics in the above-mentioned steps S901-S904 to obtain the last second timing characteristic, and no detailed description is given for this.

[0103] In one embodiment, the long short-term memory module can fuse the sixth last time series feature and the second last time series feature to obtain a fused feature, and then make a prediction based on the fused feature to output a fault result.

[0104] Among them, the fusion method can be concatenation, addition, multiplication, attention mechanism, etc., which will not be described in detail.

[0105] It should be noted that the dimension of the last second timing feature finally output by the above-mentioned multiple first timing network modules may be the same as or different from the dimension of the last sixth timing feature finally output by the multiple second timing network modules, and there is no limitation on this.

[0106] In this embodiment, since the high-dimensional features contain rich feature semantic information, and the low-dimensional features contain less feature semantic information but accurate categories. Therefore, in order to improve the prediction accuracy of the fault result, the dimension of the second last time series feature can be different from the dimension of the sixth last time series feature. Furthermore, prediction based on the feature semantic information contained in the second last time series feature and the sixth last time series feature of different dimensions can make the model prediction more accurate.

[0107] As an example, the above S904 also illustrates that the dimension of the last second time series feature is reduced to obtain the reduced dimension time series feature. Based on this, it can be considered that the dimension of the last sixth time series feature is low-dimensional, and its dimension is equal to the dimension of the fourth time series feature.

[0108] Exemplarily, the dimension of the second time series features output by the multiple first time series network modules may be 128 dimensions, and the dimension of the fifth time series features output by the multiple second time series network modules may be 64 dimensions.

[0109] Among them, the accuracy of predicting the fault results using the above long short-term memory model can be shown in the following Table 3:

[0110] Table 3:

[0111] Accuracy Recall F1-score Number of test images Lightning strike 0.92 0.90 0.92 40 Bird damage 0.91 0.86 0.84 45 foreign body 0.80 0.81 0.82 35 Wind deviation 0.81 0.83 0.81 50 Ice Cover 0.90 0.92 0.91 30

[0112] In this embodiment, the electronic device can obtain the fault recording data and meteorological data in multiple continuous cycles when the transmission line fails, and extract the features of the fault recording data and meteorological data corresponding to the multiple cycles respectively, and obtain the first fault features and first meteorological features corresponding to the multiple cycles respectively. Finally, according to the multiple first fault features and meteorological features, the fault result of the transmission line is predicted. Based on this, the fault recording data and the meteorological data that may cause the transmission line failure are combined to predict the fault result of the transmission line through multi-dimensional data features, which can improve the prediction accuracy. Moreover, the fault recording data and meteorological data are both time series data in multiple continuous cycles. Therefore, it can be considered that the corresponding multiple first fault features and first meteorological features can describe the time series feature transformation law of the fault recording data when the fault occurs. Furthermore, on the basis that the faults caused by different external factors may cause the transmission line to produce different fault recording data, the fault result is predicted based on the time series feature transformation law, which can further improve the prediction accuracy of the fault result.

[0113] In another embodiment, referring to Fig.10 , Fig.10 This is a flow chart for implementing a method for identifying a fault of a power transmission line provided by another embodiment of the present application. The electronic device can obtain fault recording data and meteorological data in multiple consecutive cycles when a fault occurs in the power transmission line. For example, the fault recording data and meteorological data in one cycle before the fault and four cycles after the fault are obtained. Then, feature extraction is performed on the fault recording data and meteorological data corresponding to the five cycles, respectively, to obtain the first fault feature and the first meteorological feature corresponding to the five cycles.

[0114] Afterwards, the first fault feature and the first meteorological feature corresponding to the first cycle can be input into the first first sequential network module to obtain the second sequential feature output by the first first sequential module. Then, the second sequential feature, as well as the first fault feature and the first meteorological feature corresponding to the second cycle are input into the second first sequential network module to obtain the second sequential feature output by the first first sequential module. This can be deduced by analogy until the last second sequential feature output by the last first sequential module is obtained.

[0115] Similar to the above method, the last sixth time series feature output by the last second time series module can be obtained. Among them, the feature dimensions output by the first time series module and the second time series module are different. For example, the dimension of the second time series feature output by the first time series module can be 128 dimensions, and the dimension of the fifth time series feature output by the second time series module can be 64 dimensions.

[0116] Finally, the sixth time series feature at the end and the second time series feature at the end are fused to obtain a fused feature, so as to make a prediction based on the fused feature and output the fault result.

[0117] See also Fig.11 , Fig.11 1 is a structural block diagram of a transmission line fault identification device provided in an embodiment of the present application. The transmission line fault identification device in this embodiment includes modules for executing Figures 1 to 10 For details, please refer to the steps in the corresponding embodiment. Figures 1 to 10 as well as Figures 1 to 10 For the convenience of explanation, only the parts related to this embodiment are shown. Fig.11 The transmission line fault identification device 1100 may include: an acquisition module 1110, a feature extraction module 1120 and a prediction module 1130, wherein:

[0118] The acquisition module 1110 is used to acquire fault recording data and meteorological data in multiple consecutive cycles when a transmission line fault occurs.

[0119] The feature extraction module 1120 is used to extract features from the fault recording data and meteorological data corresponding to multiple periods, respectively, to obtain first fault features and first meteorological features corresponding to multiple periods, respectively.

[0120] The prediction module 1130 is used to predict the fault result of the transmission line according to the multiple first fault characteristics and meteorological characteristics.

[0121] In one embodiment, the plurality of cycles includes at least one cycle before the failure and at least one cycle after the failure.

[0122] In one embodiment, the prediction module 1130 is further configured to:

[0123] A plurality of first fault characteristics and meteorological characteristics are input into a preset long short-term memory network model to obtain a fault result.

[0124] In one embodiment, the long short-term memory network model includes a plurality of first temporal network modules connected in sequence; the prediction module 1130 is further used for:

[0125] For any first sequential network module, the first target fault feature and the first meteorological feature corresponding to the first cycle are determined according to the first connection sequence of the current first sequential network module; the time sequence of the first cycle in multiple cycles is the same as the first connection sequence; the second sequential feature output by the previous first sequential network module is obtained; according to the second sequential feature, the first target fault feature and the first meteorological feature, the third sequential feature is output; the third sequential feature is the second sequential feature input to the next first sequential network module; based on the last second sequential feature output by the last first sequential network module, the fault result is predicted.

[0126] In one embodiment, the prediction module 1130 is further configured to:

[0127] The last second time series feature is reduced in dimension to obtain a fourth time series feature; and the fault result is predicted based on the fourth time series feature.

[0128] In one embodiment, the long short-term memory network model includes a plurality of second temporal network modules connected in sequence; the prediction module 1130 is further used for:

[0129] For any second sequential network module, the second target fault feature and the second meteorological feature corresponding to the second period are determined according to the second connection sequence of the current second sequential network module; the time sequence of the second period in multiple periods is the same as the second connection sequence; the fifth sequential feature output by the previous second sequential network module is obtained; the sixth sequential feature is output according to the fifth sequential feature, the second target fault feature and the second meteorological feature; the sixth sequential feature is the fifth sequential feature input to the next second sequential network module; based on the last sixth sequential feature and the last second sequential feature output by the last second sequential network module, the fault result is predicted.

[0130] In one embodiment, the prediction module 1130 is further configured to:

[0131] The sixth last time series feature and the second last time series feature are fused to obtain a fused feature; prediction is performed based on the fused feature to output the fault result.

[0132] When it is understood that Fig.11 In the structural block diagram of the transmission line fault identification device shown, each module is used to execute Figures 1 to 10 The steps in the corresponding embodiments, and for Figures 1 to 10 Each step in the corresponding embodiment has been explained in detail in the above embodiment. Figures 1 to 10 as well as Figures 1 to 10 The relevant descriptions in the corresponding embodiments are not repeated here.

[0133] Fig.12 1 is a structural block diagram of an electronic device provided by an embodiment of the present application. Fig.12 As shown, the electronic device 1200 of this embodiment includes: a processor 1210, a memory 1220, and a computer program 1230 stored in the memory 1220 and executable on the processor 1210, such as a program of a method for identifying a fault in a power transmission line. When the processor 1210 executes the computer program 1230, the steps in each embodiment of the method for identifying a fault in a power transmission line described above are implemented, such as Figure 1 Alternatively, the processor 1210 implements the above when executing the computer program 1230 Fig.11 The functions of each module in the corresponding embodiment are, for example, Fig.11 For details on the functions of each module shown, please refer to Fig.11 Related description in the corresponding embodiment.

[0134] Exemplarily, the computer program 1230 may be divided into one or more modules, one or more modules are stored in the memory 1220, and are executed by the processor 1210 to implement the power transmission line fault identification method provided in the embodiment of the present application. One or more modules may be a series of computer program instruction segments capable of completing specific functions, and the instruction segments are used to describe the execution process of the computer program 1230 in the electronic device 1200. For example, the computer program 1230 may implement the power transmission line fault identification method provided in the embodiment of the present application.

[0135] The electronic device 1200 may include, but is not limited to, a processor 1210 and a memory 1220. Those skilled in the art will appreciate that Fig.12 It is only an example of the electronic device 1200 and does not constitute a limitation of the electronic device 1200. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the electronic device may also include input and output devices, network access devices, buses, etc.

[0136] The processor 1210 may be a central processing unit, or other general-purpose processors, digital signal processors, application-specific integrated circuits, off-the-shelf programmable gate arrays or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0137] The memory 1220 may be an internal storage unit of the electronic device 1200, such as a hard disk or memory of the electronic device 1200. The memory 1220 may also be an external storage device of the electronic device 1200, such as a plug-in hard disk, a smart memory card, a flash memory card, etc. equipped on the electronic device 1200. Furthermore, the memory 1220 may also include both an internal storage unit of the electronic device 1200 and an external storage device.

[0138] An embodiment of the present application provides a computer-readable storage medium, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, a method for identifying a fault of a power transmission line as described in the above-mentioned embodiments is implemented.

[0139] An embodiment of the present application provides a computer program product. When the computer program product is run on an electronic device, the electronic device executes the power transmission line fault identification method in each of the above embodiments.

[0140] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A method for identifying a fault in a power transmission line, characterized in that: The method comprises: Obtain fault recording data and meteorological data for multiple consecutive cycles when a transmission line fault occurs; Extracting features from the fault recording data and the meteorological data corresponding to the multiple periods respectively, to obtain first fault features and first meteorological features corresponding to the multiple periods respectively; The fault result of the transmission line is predicted based on the plurality of the first fault characteristics and the meteorological characteristics.

2. The method according to claim 1, characterized in that The multiple cycles include at least at least one cycle before a fault and at least one cycle after a fault.

3. The method according to claim 1 or 2, characterized in that: The predicting the failure result of the power transmission line according to the plurality of the first fault characteristics and the meteorological characteristics comprises: Input a plurality of the first fault characteristics and the meteorological characteristics into a preset long short-term memory network model to obtain the fault result.

4. The method according to claim 3, characterized in that The long short-term memory network model includes a plurality of first sequential network modules connected in sequence; the plurality of first fault characteristics and the meteorological characteristics are input into the preset long short-term memory network model to obtain the fault result, including: For any of the first sequential network modules, determining the first target fault feature and the first meteorological feature corresponding to the first cycle according to the current first connection sequence of the first sequential network modules; the time sequence of the first cycle in the plurality of the cycles is the same as the first connection sequence; Obtaining a second timing feature output by a previous first timing network module; Outputting a third timing feature according to the second timing feature, the first target fault feature and the first meteorological feature; the third timing feature is the second timing feature input to the next first timing network module; The fault result is predicted based on the last second timing feature output by the last first timing network module.

5. The method according to claim 4, characterized in that The predicting the fault result based on the last second timing feature output by the last first timing network module includes: Performing dimension reduction on the last second time series feature to obtain a fourth time series feature; The failure consequence is predicted based on the fourth timing characteristic.

6. The method according to claim 4, characterized in that The long short-term memory network model includes a plurality of second time series network modules connected in sequence; the plurality of the first fault characteristics and the meteorological characteristics are input into the preset long short-term memory network model to obtain the fault result, and further includes: For any of the second sequential network modules, determining the second target fault characteristics and the second meteorological characteristics corresponding to the second period according to the current second connection sequence of the second sequential network modules; the time sequence of the second period in the plurality of the periods is the same as the second connection sequence; Obtaining a fifth timing feature output by a previous second timing network module; Output a sixth timing feature according to the fifth timing feature, the second target fault feature and the second meteorological feature; the sixth timing feature is the fifth timing feature input to the next second timing network module; The fault result is predicted based on the last sixth timing feature output by the last second timing network module and the last second timing feature.

7. The method according to claim 6, characterized in that The predicting the fault result based on the last sixth timing feature output by the last second timing network module and the last second timing feature includes: fusing the last sixth time series feature and the last second time series feature to obtain a fused feature; Predictions are made based on the fusion features, and the fault results are output.

8. A transmission line fault identification device, characterized in that: The device comprises: An acquisition module is used to acquire fault recording data and meteorological data in multiple consecutive cycles when a transmission line fault occurs; A feature extraction module, used to extract features from the fault recording data and the meteorological data corresponding to the multiple periods, respectively, to obtain first fault features and first meteorological features corresponding to the multiple periods respectively; A prediction module is used to predict the failure result of the transmission line according to a plurality of the first fault characteristics and the meteorological characteristics.

9. An electronic device, characterized in that: The electronic device comprises a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the electronic device implements the method as claimed in any one of claims 1 to 7.

10. A computer program product, characterized in that The invention comprises a computer program, which, when being executed, enables the method according to any one of claims 1 to 7 to be performed.

Images