Power transmission line fault type identification method and system based on multi-source information fusion
By fusion processing of multi-source information in transmission line fault data, key features are extracted and identification models are trained, the problem of low fault recognition accuracy caused by a single electrical parameter in the prior art is solved, and higher fault recognition accuracy is achieved Sex and reliability.
Patent Information
- Application Number
- CN202411695090.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-25
- Publication Date
- 2025-02-07
AI Technical Summary
The existing transmission line fault identification method uses a single electrical parameter, resulting in a small amount of feature information, low accuracy in fault identification, and insufficient use of meteorological information and reclosing information.
By fusion of meteorological information, fault recording information and reclosing information in historical fault data, key features are extracted, training samples are established using machine learning algorithms, and identification models are trained to identify fault types.
It improves the accuracy and reliability of fault identification, enables faster fault diagnosis and identification, and enhances the ability to identify multiple fault types.
Smart Images
Figure CN119395466A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of transmission line fault type identification, and more specifically, to a transmission line fault type identification method and system based on multi-source information fusion. Background Art
[0002] The increasing scale of power grids, the continuous improvement of transmission capacity and voltage levels have brought huge economic and social benefits, but at the same time, power grid failures can also cause serious harm to social economy and people's lives. Rapid and accurate grid fault classification is a prerequisite for rapid restoration of power supply, and is also an important part of fault analysis. With the continuous advancement of smart grid construction, the construction of smart distribution networks has received more and more attention from power companies, which is reflected in the continuous advancement of informationization and automation of distribution networks. In addition, modern power users have put forward higher requirements for reliable and continuous power supply.
[0003] At present, neural network models have been widely used in transmission line fault identification. Traditional transmission line fault classification methods mainly include system modeling, signal processing and artificial intelligence. Usually, the corresponding network model is obtained and trained by the electrical parameters of the transmission line, and finally the trained model is used to identify the fault; but the electrical parameters used are relatively single, resulting in less feature information extracted from the electrical parameters, resulting in lower accuracy of fault identification of the model. Summary of the invention
[0004] The purpose of the present invention is to provide a method and system for identifying the type of transmission line faults based on multi-source information fusion, which can overcome the limitations of a single data source by fusing multi-source data and extract key features, thereby making full use of multiple information to improve the accuracy of fault identification.
[0005] The above technical objectives of the present invention are achieved through the following technical solutions:
[0006] In a first aspect, the present application provides a method for identifying the type of transmission line fault based on multi-source information fusion, comprising the following specific steps:
[0007] Obtain meteorological information, fault recording information and reclosing information when each fault occurs from historical fault data, as well as the corresponding fault type each time;
[0008] Decomposing the fault recording information in the frequency domain to obtain multiple sub-frequency bands corresponding to each electrical parameter in the fault recording information, and calculating the frequency band energy of each sub-frequency band;
[0009] Calculate the index weights corresponding to the meteorological information, fault recording information and reclosing information in each fault type, and construct data groups based on the corresponding index weights according to the fault type, meteorological information, reclosing information and each frequency band energy corresponding to the fault recording information of each fault, and establish training samples through each data group;
[0010] Using each data group in the training sample to train the preset initial model until the preset initial model reaches a training end condition, and determining the preset initial model that reaches the training end condition as the recognition model;
[0011] Determine whether a fault occurs in the transmission line to be tested. When the judgment result is that a fault occurs, obtain the current line state, and input the current line state into the identification model for processing to obtain the fault type of the transmission line to be tested. The current line state includes the current meteorological state, the frequency band energy corresponding to each electrical parameter in the current recording information, and the current reclosing state.
[0012] Based on the above technical solution, the present invention can also be improved as follows.
[0013] Furthermore, the electrical parameters in the above fault recording information include at least the fault current and the zero-sequence current, the meteorological information includes at least sunny days, thunderstorm days and cloudy days, and the reclosing information includes at least successful reclosing and unsuccessful reclosing; each sub-band is specifically:
[0014] D=∑ k h k-2l D(j+1,n)+∑ k g k-2l D(j+1,n);
[0015] Where D represents one of the sub-bands, h represents the conjugate low-pass, g represents the high-pass filter coefficient, l and k both represent the position coefficient, j represents the scale coefficient, n represents the frequency index, and D(j+1, n) represents the decomposition result of the previous level.
[0016] Furthermore, the frequency band energy of each of the above sub-frequency bands is specifically:
[0017] E m，i =∑ N |D(m,i)| 2 , i=1,2,…2 m ;
[0018] In the formula, E m，i represents the frequency band energy of the ith sub-band after m times of decomposition, D(m, i) represents the ith sub-band after m times of decomposition, and N is the time length of the intercepted fault recording.
[0019] Furthermore, the above method also includes: using the energy of each frequency band of each electrical parameter in the fault recording information to perform normalization processing, constructing a fault feature vector corresponding to each electrical parameter through the energy of each frequency band after normalization processing, and constructing a data group according to the fault feature vector corresponding to each electrical parameter.
[0020] Furthermore, the above recognition model is obtained by:
[0021] Each data group is input into each classifier in the preset initial model for processing, and the recognition error of the corresponding classifier is calculated according to the output of the classifier. The recognition error is specifically:
[0022] e s =∑ r∈R V s (r)R={r|Q s (r)≠D(r)}; where e s represents the recognition error of the sth classifier, V s (r) represents the classification error of the rth sample in the sth classifier, Q s (r) represents the classification result of the rth sample given by the sth classifier, D(r) represents the actual fault type of the rth sample, and R represents the training sample;
[0023] The classifier weight is calculated based on the recognition error of the first classifier, and the classifier weight of the current classifier is used to adjust the weight of the next classifier until the recognition errors of each classifier in the preset initial model do not exceed the threshold. The preset initial model in which the recognition errors of each classifier do not exceed the threshold is determined as the recognition model, where:
[0024] The classifier weights of the current classifier, specifically:
[0025] In the formula, w s represents the classifier weight of the sth classifier, e s Represents the recognition error of the sth classifier;
[0026] The classifier weight of the current classifier is used to adjust the weight of the next classifier, specifically:
[0027] In the formula, H s represents the sample weight normalization factor, v s (r) represents the initial weight of the rth data group, and the initial weights of each data group in the training sample are the same, Q s (r) represents the classification result of the rth sample given by the sth classifier, D(r) represents the actual fault type of the rth sample, and w s+1(r) Using the classifier weight w of the current classifier s Adjust the classifier weight of the next classifier, w s represents the classifier weight of the s-th classifier.
[0028] Furthermore, the above calculations correspond to the index weights of meteorological information, fault recording information and reclosing information in each fault type, specifically:
[0029] in:
[0030] If p ij =0, then
[0031] Where W j represents the indicator weight of the jth indicator in each data group, k represents the total amount of indicators in the data group; E j represents the information entropy of the jth indicator in each data group, Y ij represents the value of the jth indicator in the i-th data group in the training sample, and n represents the total number of data groups.
[0032] Further, the above-mentioned determination of whether a fault occurs in the transmission line to be tested is specifically as follows: a target voltage of the transmission line to be tested is calculated, and when the target voltage exceeds the rated voltage of the circuit, it is determined that a fault occurs in the transmission line to be tested, and the target voltage is specifically as follows:
[0033] Where U represents the target voltage, s represents the superimposed voltage, h represents the capacitance value, M represents the ratio of the grid fault moment to the line reference angular frequency, S represents the natural frequency, L represents the length of the unloaded line, and C represents the downsampled original signal of the UHV transmission line.
[0034] In a second aspect, the present application provides a method for identifying a type of transmission line fault based on multi-source information fusion, which is applied to any one of the methods for identifying a type of transmission line fault based on multi-source information fusion in the first aspect, comprising:
[0035] The first module is used to obtain meteorological information, fault recording information and reclosing information when each fault occurs from historical fault data, as well as the corresponding fault type each time;
[0036] The second module is used to decompose the fault recording information in the frequency domain, obtain multiple sub-bands corresponding to each electrical parameter in the fault recording information, and calculate the frequency band energy of each sub-band;
[0037] The third module is used to calculate the index weights corresponding to the meteorological information, fault recording information and reclosing information in each fault type, and to construct data groups based on the corresponding index weights according to the fault type, meteorological information, reclosing information and each frequency band energy corresponding to the fault recording information of each fault, and to establish training samples through each data group;
[0038] The fourth module is used to train the preset initial model using each data group in the training sample until the preset initial model reaches the training end condition, and determine the preset initial model that reaches the training end condition as the recognition model;
[0039] The fifth module is used to determine whether a fault occurs in the transmission line to be tested. When the judgment result is that a fault occurs, the current line state is obtained, and the current line state is input into the identification model for processing to obtain the fault type of the transmission line to be tested. The current line state includes the current meteorological state, the frequency band energy corresponding to each electrical parameter in the current recording information, and the current reclosing state.
[0040] In a third aspect, the present application provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements any one of the methods in the first aspect when executing the computer program.
[0041] In a fourth aspect, the present application provides a non-transitory computer-readable storage medium, which stores computer instructions, and the computer instructions enable a computer to execute any one of the methods in the first aspect.
[0042] Compared with the prior art, the present invention has at least the following beneficial effects:
[0043] In the present application, since the current transmission line fault diagnosis involves little fault information, most of them only focus on a single electrical parameter, while environmental meteorological information and reclosing information are not taken into consideration; single fault information will lead to diagnostic errors due to the small amount of characteristic information; in the present application, firstly, the meteorological information, fault recording information and reclosing information at each fault occurrence in the historical fault data, as well as the corresponding fault type each time are used as data samples, and the fault recording information includes information on multiple electrical parameters, such as fault current and zero-sequence current, so as to extract features of the fault current and zero-sequence current, that is, decompose them in the frequency domain to obtain multiple sub-bands corresponding to each electrical parameter in the fault recording information, and calculate the band energy of each sub-band; secondly, by calculating the fault type each time The index weights corresponding to the meteorological information, fault recording information and reclosing information are used to express the influence of meteorological information, fault recording information and reclosing information in each fault type, so that the relationship between the fault type and each state information can be more accurately described when training the model; so far, based on the corresponding index weights, the fault type, meteorological information, reclosing information of each fault, and the energy of each frequency band corresponding to the fault recording information are used to establish training samples, and the preset initial model is trained through the established training samples; finally, when the judgment result of the transmission line to be tested is that a fault has occurred, the current line state is obtained, and the current line state is processed by the recognition model obtained through training, so as to obtain the fault type of the transmission line to be tested.
[0044] In the present application, multiple electrical parameters in the fault recording information are decomposed in the frequency domain, and the band energy of each sub-band obtained after the decomposition is calculated to extract the fault characteristic values of different transmission line fault types. Multi-source information fusion combines meteorological information and reclosing information with the fault characteristic values of the extracted electrical parameters, and calculates the indicator weights corresponding to the meteorological information, fault recording information and reclosing information in each fault type to express the influence of meteorological information, fault recording information and reclosing information in each fault type, and then uses a machine learning algorithm to complete the fault diagnosis of multi-source information fusion; in this way, in terms of weight calculation and multi-source information fusion, the relationship between the fault type and each state information can be more accurately described when training the model, so that the model obtained after training has a more accurate recognition rate, higher accuracy and reliability, and can also achieve faster fault diagnosis and identification when used. BRIEF DESCRIPTION OF THE DRAWINGS
[0045] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, constitute a part of this application, and do not constitute a limitation of the embodiments of the present invention. In the drawings:
[0046] Figure 1 A method flow chart of an identification method in an embodiment of the present invention;
[0047] Figure 2 A connection diagram of an identification system in an embodiment of the present invention;
[0048] Figure 3 Schematic diagram of the connection of electronic equipment in an embodiment of the present invention. DETAILED DESCRIPTION
[0049] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.
[0050] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0051] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.
[0052] In the description of the embodiments of the present invention, "plurality" means at least 2.
[0053] Embodiment 1:
[0054] In order to overcome the limitation of a single data source, by fusing multi-source data, extracting key features and making full use of multiple information to improve the accuracy of fault identification, this embodiment provides a transmission line fault type identification method based on multi-source information fusion, such as Figure 1 As shown, the specific steps include:
[0055] S1, obtain the meteorological information, fault recording information and reclosing information when each fault occurs from the historical fault data, as well as the corresponding fault type each time.
[0056] Specifically, the electrical parameters in the above-mentioned fault recording information include at least the fault current and the zero-sequence current, the meteorological information includes at least sunny days, thunderstorm days and cloudy days, and the reclosing information includes at least successful reclosing and unsuccessful reclosing.
[0057] Among them, the fault types may include lightning strikes, external damage, bird damage, wildfires, grounding, etc., and the data groups corresponding to these fault types include meteorological information, fault recording information and reclosing information when the fault occurs.
[0058] S2, decomposing the fault recording information in the frequency domain to obtain a plurality of sub-frequency bands corresponding to the electrical parameters in the fault recording information, and calculating the frequency band energy of each sub-frequency band.
[0059] Among them, since the time domain characteristics of the fault current and the zero-sequence current are difficult to reflect the fault characteristics, in this embodiment, the fault characteristic quantities are extracted from the frequency domain perspective; specifically, each sub-frequency band is specifically:
[0060] D=∑ k h k-2l D(j+1,n)+∑ k g k-2l D(j+1,n);
[0061] Where D represents one of the sub-bands, h represents the conjugate low-pass, g represents the high-pass filter coefficient, l and k both represent the position coefficient, j represents the scale coefficient, n represents the frequency index, and D(j+1, n) represents the decomposition result of the previous level.
[0062] Optionally, the frequency band energy of each of the above sub-frequency bands is specifically:
[0063] E m，i =∑ N |D(m,i)| 2 , i = 1, 2, ... 2 m ;
[0064] In the formula, E m，i represents the frequency band energy of the ith sub-band after m times of decomposition, D(m, i) represents the ith sub-band after m times of decomposition, and N is the time length of the intercepted fault recording.
[0065] Among them, when the fault current or zero-sequence current in the fault recording information is decomposed in the frequency domain, two sub-bands can be obtained by decomposing it once, four sub-bands can be obtained by decomposing it twice, and eight sub-bands can be obtained by decomposing it three times. That is, the number of sub-bands obtained is twice the number of decompositions. m .
[0066] Optionally, the above method also includes: using the energy of each frequency band of each electrical parameter in the fault recording information to perform normalization processing, constructing a fault feature vector corresponding to each electrical parameter through the normalized energy of each frequency band, and constructing a data group according to the fault feature vector corresponding to each electrical parameter.
[0067] Each fault characteristic vector includes the frequency band energy corresponding to each sub-frequency band in the fault current or zero-sequence current.
[0068] S3, calculate the index weights corresponding to the meteorological information, fault recording information and reclosing information in each fault type, and construct data groups based on the corresponding index weights with the fault type, meteorological information, reclosing information and each frequency band energy corresponding to the fault recording information of each fault, and establish training samples through each data group.
[0069] Optionally, the above calculation of the index weights corresponding to the meteorological information, fault recording information and reclosing information in each fault type is specifically:
[0070] in:
[0071] If p ij =0, then
[0072] Where W j represents the indicator weight of the jth indicator in each data group, k represents the total amount of indicators in the data group; E j represents the information entropy of the jth indicator in each data group, Y ij represents the value of the jth indicator in the i-th data group in the training sample, and n represents the total number of data groups.
[0073] Among them, by calculating the weight of each indicator in the data group, the fault type can be associated with the meteorological information, frequency band energy and reclosing information again. After calculating the indicator weight, each data group can be expressed as: the fault type is lightning strike, the indicator weight of meteorological information can be 0.5, the indicator weight of frequency band energy can be 0.25, and the indicator weight of reclosing information is 0.25. That is, for the same data group, the sum of the weights of each indicator is 1.
[0074] Specifically, the influence of meteorological information, fault recording information and reclosing information in each fault type is expressed by calculating the indicator weights corresponding to each fault type. This allows a more accurate description of the relationship between the fault type and each state information when training the model.
[0075] S4, using each data group in the training sample to train the preset initial model until the preset initial model reaches the training end condition, and determining the preset initial model that reaches the training end condition as the recognition model.
[0076] Optionally, the above recognition model is obtained by:
[0077] S41, input each data group into each classifier in the preset initial model for processing, and calculate the recognition error of the corresponding classifier according to the output of the classifier. The recognition error is specifically:
[0078] e s =∑ r∈R V s (r)R={r|Q s (r)≠D(r)}; where e s represents the recognition error of the sth classifier, V s (r) represents the classification error of the rth sample in the sth classifier, Q s (r) represents the classification result of the rth sample given by the sth classifier, D(r) represents the actual fault type of the rth sample, and R represents the training sample.
[0079] S42, calculating the classifier weight according to the recognition error of the first classifier, and adjusting the weight of the next classifier by using the classifier weight of the current classifier, until the recognition errors of each classifier in the preset initial model do not exceed the threshold, and determining the preset initial model in which the recognition errors of each classifier do not exceed the threshold as the recognition model, wherein:
[0080] The classifier weights of the current classifier, specifically:
[0081] In the formula, w s represents the classifier weight of the sth classifier, e s Represents the recognition error of the sth classifier;
[0082] Furthermore, the classifier weight of the current classifier is used to adjust the weight of the next classifier, specifically:
[0083] In the formula, H s represents the sample weight normalization factor, v s (r) represents the initial weight of the rth data group, and the initial weights of each data group in the training sample are the same, Q s (r) represents the classification result of the rth sample given by the sth classifier, D(r) represents the actual fault type of the rth sample, and w s+1 (r) Using the classifier weight w of the current classifier s Adjust the classifier weight of the next classifier, w s represents the classifier weight of the s-th classifier.
[0084] S5, determine whether a fault occurs in the transmission line to be tested. When the judgment result is that a fault occurs, obtain the current line state, and input the current line state into the identification model for processing to obtain the fault type of the transmission line to be tested. The current line state includes the current meteorological state, the frequency band energy corresponding to each electrical parameter in the current recording information, and the current reclosing state.
[0085] Among them, when the recognition model training is completed and put into use, it is necessary to monitor in real time whether the transmission line to be tested has a fault, which can be judged by calculating the target voltage.
[0086] Optionally, the above-mentioned determination of whether a fault occurs in the transmission line to be tested is specifically: calculating a target voltage of the transmission line to be tested, and when the target voltage exceeds the rated voltage of the circuit, determining that the transmission line to be tested has a fault, the target voltage is specifically:
[0087] Where U represents the target voltage, s represents the superimposed voltage, h represents the capacitance value, M represents the ratio of the grid fault moment to the line reference angular frequency, S represents the natural frequency, L represents the length of the unloaded line, and C represents the downsampled original signal of the UHV transmission line.
[0088] Embodiment 2:
[0089] The embodiment of the present application provides a method for identifying a type of transmission line fault based on multi-source information fusion, which is applied to any one of the methods for identifying a type of transmission line fault based on multi-source information fusion in Embodiment 1, such as Figure 2 As shown, including:
[0090] The first module is used to obtain meteorological information, fault recording information and reclosing information when each fault occurs from historical fault data, as well as the corresponding fault type each time.
[0091] The second module is used to decompose the fault recording information in the frequency domain to obtain a plurality of sub-frequency bands corresponding to each electrical parameter in the fault recording information, and calculate the frequency band energy of each sub-frequency band.
[0092] The third module is used to calculate the index weights corresponding to the meteorological information, fault recording information and reclosing information in each fault type, and construct data groups based on the corresponding index weights with the fault type, meteorological information, reclosing information and each frequency band energy corresponding to the fault recording information of each fault, and establish training samples through each data group.
[0093] The fourth module is used to train the preset initial model using each data group in the training sample until the preset initial model reaches the training end condition, and determine the preset initial model that reaches the training end condition as the recognition model.
[0094] The fifth module is used to determine whether a fault occurs in the transmission line to be tested. When the judgment result is that a fault occurs, the current line state is obtained, and the current line state is input into the identification model for processing to obtain the fault type of the transmission line to be tested. The current line state includes the current meteorological state, the frequency band energy corresponding to each electrical parameter in the current recording information, and the current reclosing state.
[0095] Specifically, firstly, the meteorological information, fault recording information and reclosing information at each fault occurrence in the historical fault data, as well as the corresponding fault type each time, are taken as data samples, and the fault recording information includes information of multiple electrical parameters, such as fault current and zero-sequence current, so as to extract the features of the fault current and zero-sequence current, that is, decompose them in the frequency domain to obtain multiple sub-frequency bands corresponding to each electrical parameter in the fault recording information, and calculate the frequency band energy of each sub-frequency band; secondly, the influence of the meteorological information in each fault type is expressed by calculating the index weights corresponding to the meteorological information, fault recording information and reclosing information in each fault type. The influence degree of the fault recording information and the influence degree of the reclosing information are determined, so that the relationship between the fault type and each state information can be more accurately described when training the model; thus, based on the corresponding indicator weights, the fault type, meteorological information, reclosing information of each fault, and the energy of each frequency band corresponding to the fault recording information are used to establish training samples, and the preset initial model is trained through the established training samples; finally, when the judgment result of the transmission line to be tested is that a fault occurs, the current line state is obtained, and the current line state is processed by the recognition model obtained through the training, so as to obtain the fault type of the transmission line to be tested.
[0096] Embodiment 3:
[0097] The present application embodiment provides an electronic device, such as Figure 3 As shown, it includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, any method in Embodiment 1 is implemented.
[0098] Embodiment 4:
[0099] An embodiment of the present application provides a non-transitory computer-readable storage medium, which stores computer instructions, and the computer instructions enable a computer to execute any method in Example 1.
[0100] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.
Claims
1. A method for identifying transmission line fault types based on multi-source information fusion, characterized in that: The specific steps include: Obtain meteorological information, fault recording information and reclosing information when each fault occurs from historical fault data, as well as the corresponding fault type each time; Decomposing the fault recording information in the frequency domain to obtain a plurality of sub-frequency bands corresponding to the electrical parameters in the fault recording information, and calculating the frequency band energy of each sub-frequency band; Calculate the index weights corresponding to the meteorological information, the fault recording information and the reclosing information in each fault type, respectively, and construct data groups based on the corresponding index weights with the fault type, the meteorological information, the reclosing information and the energy of each frequency band corresponding to the fault recording information of each fault, and establish training samples through each data group; Using each data group in the training sample to train the preset initial model until the preset initial model reaches a training end condition, and determining the preset initial model that reaches the training end condition as the recognition model; Determine whether a fault occurs in the transmission line to be tested. When the judgment result is that a fault occurs, obtain the current line state, and input the current line state into the identification model for processing to obtain the fault type of the transmission line to be tested. The current line state includes the current meteorological state, the frequency band energy corresponding to each electrical parameter in the current recording information, and the current reclosing state.
2. The method for identifying the type of transmission line fault based on multi-source information fusion according to claim 1 is characterized in that: The electrical parameters in the fault recording information include at least the fault current and the zero-sequence current, the meteorological information includes at least sunny days, thunderstorm days and cloudy days, and the reclosing information includes at least successful reclosing and unsuccessful reclosing; each of the sub-frequency bands is specifically: D=∑ k h k-2l D(j+1,n)+∑ k g k-2l D(j+1,n); Where D represents one of the sub-bands, h represents the conjugate low-pass, g represents the high-pass filter coefficient, l and k both represent the position coefficient, j represents the scale coefficient, n represents the frequency index, and D(j+1, n) represents the decomposition result of the previous level.
3. The method for identifying the type of transmission line fault based on multi-source information fusion according to claim 1 is characterized in that: The frequency band energy of each sub-band is specifically: E m，i =∑ N |D(m,i)| 2 ,i=1,2,…2 m ; In the formula, E m，i represents the frequency band energy of the ith sub-band after m times of decomposition, D(m, i) represents the ith sub-band after m times of decomposition, and N is the time length of the intercepted fault recording.
4. The method for identifying transmission line fault types based on multi-source information fusion according to claim 1, characterized in that: The method also includes: performing normalization processing using the energy of each frequency band of each electrical parameter in the fault recording information, constructing a fault feature vector corresponding to each electrical parameter using the energy of each frequency band after the normalization processing, and constructing the data group according to the fault feature vector corresponding to each electrical parameter.
5. The method for identifying transmission line fault types based on multi-source information fusion according to claim 1, characterized in that: The recognition model is obtained in the following way: Each of the data groups is input into each classifier in the preset initial model for processing, and the recognition error of the corresponding classifier is calculated according to the output of the classifier. The recognition error is specifically: e s =∑ r∈R V s (r)R={r|Q s (r)≠D(r)}; where e s represents the recognition error of the sth classifier, V s (r) represents the classification error of the rth sample in the sth classifier, Q s (r) represents the classification result of the r-th sample given by the s-th classifier, D(r) represents the actual fault type of the r-th sample, and R represents the training sample; The classifier weight is calculated based on the recognition error of the first classifier, and the classifier weight of the current classifier is used to adjust the weight of the next classifier until the recognition errors of each classifier in the preset initial model do not exceed the threshold. The preset initial model in which the recognition errors of each classifier do not exceed the threshold is determined as the recognition model, where: The classifier weight of the current classifier is specifically: In the formula, w s represents the classifier weight of the sth classifier, e s represents the recognition error of the sth classifier; The weight of the next classifier is adjusted by using the classifier weight of the current classifier, specifically: In the formula, H s represents the sample weight normalization factor, v s (r) represents the initial weight of the rth data group, and the initial weights of each data group in the training sample are the same, Q s (r) represents the classification result of the rth sample given by the sth classifier, D(r) represents the actual fault type of the rth sample, and w s+1 (r) Using the classifier weight w of the current classifier s Adjust the classifier weight of the next classifier, w s represents the classifier weight of the s-th classifier.
6. The method for identifying the type of transmission line fault based on multi-source information fusion according to claim 5, characterized in that: The calculation of the index weights corresponding to the meteorological information, fault recording information and reclosing information in each fault type is specifically as follows: in: If p ij =0, then Where W j represents the indicator weight of the jth indicator in each data group, k represents the total amount of indicators in the data group; E j represents the information entropy of the jth indicator in each data group, Y ij represents the value of the jth indicator in the i-th data group in the training sample, and n represents the total number of data groups.
7. The method for identifying the type of transmission line fault based on multi-source information fusion according to claim 1, characterized in that: The determining whether a fault occurs in the transmission line to be tested is specifically: calculating a target voltage of the transmission line to be tested, and when the target voltage exceeds the rated voltage of the circuit, determining that a fault occurs in the transmission line to be tested, wherein the target voltage is specifically: Where U represents the target voltage, s represents the superimposed voltage, h represents the capacitance value, M represents the ratio of the grid fault moment to the line reference angular frequency, S represents the natural frequency, L represents the length of the unloaded line, and C represents the downsampled original signal of the UHV transmission line.
8. A method for identifying the type of fault of a power transmission line based on multi-source information fusion, applied to the method for identifying the type of fault of a power transmission line based on multi-source information fusion according to any one of claims 1 to 7, characterized in that: include: The first module is used to obtain meteorological information, fault recording information and reclosing information when each fault occurs from historical fault data, as well as the corresponding fault type each time; The second module is used to decompose the fault recording information in the frequency domain to obtain multiple sub-frequency bands corresponding to each electrical parameter in the fault recording information, and calculate the frequency band energy of each sub-frequency band; The third module is used to calculate the index weights corresponding to the meteorological information, the fault recording information and the reclosing information in each fault type, and to construct a data group based on the corresponding index weights with the fault type, the meteorological information, the reclosing information and the energy of each frequency band corresponding to the fault recording information of each fault, and to establish a training sample through each data group; The fourth module is used to train the preset initial model using each data group in the training sample until the preset initial model reaches the training end condition, and determine the preset initial model that reaches the training end condition as the recognition model; The fifth module is used to determine whether a fault occurs in the transmission line to be tested. When the judgment result is that a fault occurs, the current line state is obtained, and the current line state is input into the identification model for processing to obtain the fault type of the transmission line to be tested. The current line state includes the current meteorological state, the frequency band energy corresponding to each electrical parameter in the current recording information, and the current reclosing state.
9. An electronic device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the method according to any one of claims 1 to 7 is implemented when the processor executes the computer program.
10. A non-transitory computer-readable storage medium, characterized in that: The non-transitory computer-readable storage medium stores computer instructions, which enable a computer to execute the method of any one of claims 1-7.