Method, device and computer equipment for determining fault cause of power transmission line

Abstract

The present application relates to a method, device, computer equipment and storage medium for determining the cause of a transmission line fault. The method comprises: obtaining transmission line fault recording data and meteorological data, and then extracting fault features, and determining the cause of the transmission line fault according to the fault features and a weighted naive Bayesian model. In this embodiment, the weighted naive Bayesian model fault features are used for analysis to determine the cause of the transmission line fault. The use of the weighted naive Bayesian model to determine the cause of the transmission line fault can improve the real-time nature of the fault cause analysis, and there is no need for power grid line patrol personnel to patrol the suspected fault area located by ranging, which can effectively improve the efficiency of fault cause determination. Moreover, there is no need for power grid line patrol personnel to patrol the suspected fault area located by ranging, which also saves labor costs.

Description

Technical Field

[0001] The present application relates to the field of power grid technology, and in particular to a method, device, computer equipment and storage medium for determining the cause of a transmission line fault. Background Art

[0002] With the rapid development of my country's economy and the continuous progress of the power industry, the scale, capacity and coverage of modern power grids are getting larger and larger. The power grid occupies an important position in my country's national economy and people's living standards, and is spread throughout all aspects of people's lives. Therefore, power outages caused by faults will cause huge losses to social production and people's lives.

[0003] Due to the uneven distribution of power loads in various regions, it is often necessary to use transmission lines to achieve long-distance, large-capacity power transmission. As an important part of the power system, transmission lines play a key role in the safety and stability of power transmission. However, they are widely distributed and have a complex and changeable operating environment. They are easily affected by severe weather such as lightning, strong winds, ice and snow, or external damage, which can cause tripping and cause great inconvenience to people's lives. At present, the fault diagnosis of transmission lines mainly relies on the grid patrol personnel to gradually check the suspected faults located by ranging, and then organize relevant personnel to carry out emergency repairs.

[0004] However, the existing method of manually troubleshooting power transmission circuit faults has the problem of low efficiency. Summary of the invention

[0005] Based on this, it is necessary to provide a method, device, computer equipment and storage medium for determining the cause of a transmission line fault, which can improve the efficiency of determining the cause of a transmission line fault, in order to address the above technical problems.

[0006] A method for determining a fault cause of a power transmission line, the method comprising:

[0007] Obtain fault recording data and meteorological data of the power grid system;

[0008] Extracting features from the fault recording data and the meteorological data to obtain fault features of the power transmission line of the power grid system;

[0009] According to the fault characteristics and a preset weighted naive Bayesian model, a target fault cause of a transmission line in the power grid system is determined; the weighted naive Bayesian model is obtained according to a weighted naive Bayesian classification algorithm based on a composite algorithm.

[0010] In one embodiment, the weighted naive Bayes model preset according to the fault characteristics determines the fault cause of the transmission line in the power grid system, including:

[0011] Determine a target fault feature corresponding to each fault cause according to the correspondence between the fault cause and the fault feature and the fault feature of the transmission line;

[0012] Inputting the target fault feature corresponding to each of the fault causes into the weighted naive Bayes model to obtain the classification error rate of each of the fault causes;

[0013] The fault cause corresponding to the minimum classification error rate is determined as the target fault cause of the transmission line.

[0014] In one embodiment, the weighted naive Bayes model includes a weight of each fault cause, and the weight of each fault cause is determined by a composite algorithm.

[0015] In one embodiment, the training method of the weighted naive Bayes model includes:

[0016] Acquire a fault feature sample set of a transmission line; the fault feature sample set includes sample fault features and actual fault causes corresponding to the sample fault features;

[0017] The preset initial naive Bayesian model is trained according to the fault feature sample set based on a composite algorithm to obtain the weighted naive Bayesian model.

[0018] In one embodiment, the composite algorithm is used to train a preset initial naive Bayesian model according to the fault feature sample set to obtain the weighted naive Bayesian model, including:

[0019] Randomly extracting a preset number of fault feature samples from the fault feature sample set as a first training set, and using the remaining fault feature samples as a second training set;

[0020] Training the initial naive Bayes model according to the first training set to obtain a naive Bayes model;

[0021] The naive Bayes model is trained according to the second training set based on the composite algorithm to obtain the weighted naive Bayes model.

[0022] In one embodiment, the step of training the naive Bayesian model based on the composite algorithm according to the second training set to obtain the weighted naive Bayesian model includes:

[0023] Constructing an initial composite shape according to the sample fault features in the second training set;

[0024] Determining the weight of the best point and the weight of the worst point according to the objective function value of each vertex in the initial composite shape;

[0025] Calculate the reflection point according to the weight of the worst point and the weight of the center point; the center point is the center point of other vertices except the worst point;

[0026] According to the objective function value of the reflection point and the objective function value of the worst point, an iterative operation is performed until an iteration termination condition is met, and the weight of the best point that meets the iteration termination condition is determined as the weight of the fault cause in the naive Bayes model to obtain the weighted naive Bayes model.

[0027] In one embodiment, performing iterative calculation according to the objective function value of the reflection point and the objective function value of the worst point includes:

[0028] Determining whether the reflection point is a feasible point;

[0029] If the reflection point is a feasible point and the objective function value of the reflection point is less than the objective function value of the worst point, the reflection point is used to replace the worst point to obtain a new composite shape, and the step of determining the weight of the best point and the weight of the worst point according to the objective function value of each vertex in the new composite shape is returned;

[0030] If the reflection point is a feasible point and the objective function value of the reflection point is not less than the objective function value of the worst point, the reflection point is moved closer to the center point, and a new reflection point is calculated until the objective function value of the new reflection point is less than the objective function value of the worst point;

[0031] If the reflection point is not a feasible point, the reflection point moves closer to the center point until the new reflection point becomes a feasible point, and then the step of making the reflection point a feasible point is performed.

[0032] In one embodiment, the iteration termination condition includes that the root mean square value of the difference between the objective function value of each vertex of the composite shape and the objective function value of the best point is less than a preset accuracy value.

[0033] A device for determining the cause of a power transmission line fault, the device comprising:

[0034] An acquisition module is used to acquire fault recording data and meteorological data of the power grid system;

[0035] An extraction module, used for extracting features from the fault recording data and the meteorological data to obtain fault features of the power transmission line of the power grid system;

[0036] A determination module is used to determine the target fault cause of the transmission line in the power grid system according to the fault characteristics and a preset weighted naive Bayes model; the weighted naive Bayes model is obtained according to a weighted naive Bayes classification algorithm based on a composite algorithm.

[0037] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0038] Obtain fault recording data and meteorological data of the power grid system;

[0039] Extracting features from the fault recording data and the meteorological data to obtain fault features of the power transmission line of the power grid system;

[0040] According to the fault characteristics and a preset weighted naive Bayesian model, a target fault cause of a transmission line in the power grid system is determined; the weighted naive Bayesian model is obtained according to a weighted naive Bayesian classification algorithm based on a composite algorithm.

[0041] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the following steps:

[0042] Obtain fault recording data and meteorological data of the power grid system;

[0043] Extracting features from the fault recording data and the meteorological data to obtain fault features of the power transmission line of the power grid system;

[0044] According to the fault characteristics and a preset weighted naive Bayesian model, a target fault cause of a transmission line in the power grid system is determined; the weighted naive Bayesian model is obtained according to a weighted naive Bayesian classification algorithm based on a composite algorithm.

[0045] The above-mentioned method, device, computer equipment and storage medium for determining the cause of the transmission line fault first obtains the transmission line fault recording data and meteorological data, then extracts the fault characteristics, and finally inputs the fault characteristics into the weighted naive Bayes model to determine the cause of the transmission line fault. The data is mainly obtained from the power grid system and the meteorological system, the acquisition method is simple, and the existing historical data and engineering experience have a certain guidance for the extraction of features. When the weighted naive Bayes model is used to determine the cause of the transmission line fault, the acquired fault characteristics are directly analyzed, and there is no need for the power grid patrol personnel to patrol the suspected fault area located by ranging. Therefore, this embodiment can effectively improve the efficiency of determining the cause of the transmission line fault. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 is a block diagram of a computer device in one embodiment;

[0047] Figure 2 A schematic flow chart of a method for determining a fault cause of a power transmission line in one embodiment;

[0048] Figure 3 A flowchart of a specific process of determining the cause of a power transmission line failure in one embodiment;

[0049] Figure 4 A schematic flow chart of a method for training the causes of transmission line faults in one embodiment;

[0050] Figure 5 A schematic diagram of a process of weighted naive Bayes training based on a composite algorithm in one embodiment;

[0051] Figure 6 A schematic diagram of a process of training the initial naive Bayes based on the second training set based on a composite algorithm in one embodiment;

[0052] Figure 7 A schematic diagram of a flow chart of iterative operation in one embodiment;

[0053] Figure 8 is a schematic diagram of a flow chart of iterative operation in another embodiment;

[0054] Fig. 9 A structural block diagram of a device for determining the cause of a fault of a power transmission line in one embodiment;

[0055] Fig.10 A structural block diagram of a device for determining the cause of a power transmission line fault in another embodiment

[0056] Fig.11 A structural block diagram of a device for determining a fault cause of a power transmission line in another embodiment;

[0057] Fig.12 A structural block diagram of a device for determining the cause of a power transmission line fault in another embodiment

[0058] Fig.13 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0060] The present application provides a method for determining the cause of a transmission line fault, which can be applied to a power grid system application environment, and the application environment includes a server. The server obtains fault recording data from the power system and meteorological data from the meteorological system through the network, extracts features from the fault recording data and meteorological data, and inputs the extracted fault features into a weighted naive Bayes model to obtain the cause of the transmission line fault in the power grid system. The server can be implemented as an independent server or a server cluster consisting of multiple servers.

[0061] In one embodiment, a computer device that can be used as the server is provided. The computer device can be a terminal, and its internal structure diagram can be as follows: Figure 1 As shown. The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a method for determining the cause of a power transmission line fault is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a key, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0062] In one embodiment, Figure 2 As shown, a method for determining the cause of a transmission line fault is provided, and the method is applied to Figure 1 Taking the server shown in the figure as an example, the method includes the following steps:

[0063] S201, acquiring fault recording data and meteorological data of the power grid system.

[0064] Among them, fault recording data refers to the changes in relevant electrical quantities and the action behaviors of relay protection and automatic safety devices after a large disturbance occurs in the power grid system, including the transition resistance value, transition voltage-ampere characteristics, zero-sequence current DC content, zero-sequence current harmonic content and reclosing characteristics of the transmission lines and equipment in the power grid system; meteorological data mainly refers to climate data and weather data, including weather, time period, season, etc.

[0065] In this embodiment, the server can obtain the fault recording data from the database of the power grid system and the meteorological data from the database of the meteorological station. The server can obtain the fault recording data and meteorological data periodically, in real time, or after receiving a user instruction.

[0066] S202, extracting features from the fault recording data and meteorological data to obtain fault features of the transmission lines of the power grid system.

[0067] In this embodiment, when extracting features from fault recording data and meteorological data, effective fault features can be selected according to the following two basic principles:

[0068] Principle 1: Fault characteristics must be closely related to the fault cause type, and each fault characteristic must be independent of each other.

[0069] Principle 2: The number of fault features must be able to be effectively extracted and calculated, and the fault recording data and meteorological data used for fault feature extraction can be obtained promptly and accurately after the fault.

[0070] The above principles are used to obtain fault recording data and meteorological data, but the collected fault recording data and meteorological data have various forms of expression, and the fault characteristics need to be summarized and sorted. For example, the characteristic values ​​of the time period characteristics can be obtained periodically every five minutes or in real time during the collection process. When a fault occurs in the power transmission line of the power grid system, the values ​​of the fault characteristics are divided according to Table 1 and summarized into different categories.

[0071] Table 1 Transmission line fault characteristics and values

[0072]

[0073] In this embodiment, keywords can be used to extract features from fault recording data and meteorological data to obtain fault features of the power transmission line of the power grid system. For example, the eight features in Table 1 are used as keywords to extract feature values ​​corresponding to each keyword in the fault recording data and meteorological data. Alternatively, a neural network model can be used to extract fault features, and the fault recording data and meteorological data are input into the neural network model to directly output the fault features. This is not limited in the embodiments of the present application.

[0074] S203, determining the target fault cause of the transmission line in the power grid system according to the fault characteristics and a preset weighted naive Bayes model; the weighted naive Bayes model is obtained according to a weighted naive Bayes classification algorithm based on a composite algorithm.

[0075] Among them, the naive Bayes classification model is a simple and efficient classification model, but it must meet the attribute independence assumption, which directly affects its classification performance. In order to weaken the assumption problem of attribute independence, the weighted naive Bayes classification model improves the naive Bayes classification model, which can include the following derivation process:

[0076] Let D be a set of training tuples and class labels, and each tuple is represented by an n-dimensional vector representing the attribute vector, such as X = {x1, x2, ..., x n}, describing n attributes X1, X2, ... X n n measurements of tuples. Assume there are m categories, C1, C2, …C m Given a training tuple X, predict that X belongs to category C i , if and only if:

[0077] P(C i |X)>P(C j |X)1≤j≤m,j≠m (1)

[0078] The posterior probability P(C i |X) is calculated as:

[0079]

[0080] When P(C i |X) is maximized, and the category at this time is called C i is the maximum a posteriori hypothesis.

[0081] Class prior probability P(C i ) can be used with P(C i )=|C i,D | / |D| estimate, where |C i ,D|is C in D i The number of training tuples of the class. Naive Bayes assumes that each attribute is conditionally independent, that is,

[0082]

[0083] According to whether the attribute value is discrete or continuous, the probability value p(x k |C i )’s estimate:

[0084] If X k is a discrete value, then p(x k |C i ) is attribute X k The median value is x k and belongs to category C i The number of tuples divided by the category C i The number of tuples of |C i,D|.

[0085] If Xk is continuous valued, it is assumed that the continuous valued attributes follow a Gaussian distribution:

[0086]

[0087] therefore:

[0088]

[0089] The second and third parameters μ and σ in the function g are C i Class training tuple attribute X k The mean and standard deviation of .

[0090] Since P(X) is a constant, equation (2) is modified to the following equation, which is called the naive Bayes classification model, C NB (X) represents the accuracy of fault classification of the naive Bayes model when the attribute is X:

[0091]

[0092] Since it is difficult to satisfy the assumption of conditional independence of Naive Bayes in practice, different attributes are given different weights according to their classification importance, so that Naive Bayes can be expanded. The weighted Naive Bayes model is as follows:

[0093]

[0094] In the formula, w k Represents attribute X k The weight, C WNB (X) represents the accuracy of fault classification of the weighted naive Bayes model when the attribute is X.

[0095] In this embodiment, a composite method may be used to determine the weights of the attributes of the weighted naive Bayes model, thereby obtaining an optimized weighted naive Bayes model.

[0096] In this embodiment, the fault feature can be input into the weighted naive Bayesian model to obtain the classification error rate corresponding to each fault feature, and the fault cause can be determined according to the classification error rate corresponding to each fault feature. For example, the corresponding relationship between the fault feature and the fault cause can be established in advance, but when the classification error rate of each fault feature is obtained through the weighted naive Bayesian model, the fault cause corresponding to the fault feature with the lowest classification error rate can be determined as the target fault cause. Alternatively, multiple fault features can be classified and combined to obtain the fault cause corresponding to each fault feature combination, as shown in Table 2, which gives the corresponding relationship between the fault feature and the fault cause. After the fault feature is obtained, the fault feature can be combined in the manner shown in Table 2, and then each fault feature combination is input into the weighted naive Bayesian model to obtain the classification error rate corresponding to each fault feature combination, and the fault cause corresponding to the fault feature combination with the lowest classification error rate is determined as the target fault cause. Alternatively, each fault feature may be input into the weighted naive Bayes model to obtain the classification error rate of each fault feature, and then the average classification error rate corresponding to each fault feature combination is calculated according to Table 2, and the fault cause corresponding to the fault feature combination with the lowest average classification error rate is determined as the target fault cause.

[0097] Table 2 Correspondence between fault characteristics and fault causes

[0098]

[0099] It should be noted that the correspondence between the above-mentioned fault characteristics and fault causes does not mean that there is a unique correspondence between the fault cause and the fault characteristics. It only means that under certain conditions, the corresponding type of fault cause is more likely to occur, and the measurement of this possibility needs to be based on the analysis of historical actual fault data.

[0100] The method for determining the cause of a transmission line fault provided in an embodiment of the present application obtains transmission line fault recording data and meteorological data, and then extracts fault features, and determines the cause of the transmission line fault based on the fault features and the weighted naive Bayes model. In this embodiment, the weighted naive Bayes model fault features are used for analysis to determine the cause of the transmission line fault. The use of the weighted naive Bayes model to determine the cause of the transmission line fault can improve the real-time nature of the fault cause analysis, and does not require power grid line patrol personnel to patrol the suspected fault area located by ranging, which can effectively improve the efficiency of fault cause determination. Moreover, there is no need for power grid line patrol personnel to patrol the suspected fault area located by ranging, which also saves labor costs.

[0101] exist Figure 2 In the embodiment shown, a method for determining the cause of a transmission line fault based on a weighted naive Bayes model is introduced. Figure 3Taking the embodiment as an example, the specific process of determining the fault cause of the power transmission line in the power grid system by using the weighted naive Bayes model preset according to the fault characteristics is mainly introduced. Figure 3 As shown, the following steps are included:

[0102] S301, determining a target fault feature corresponding to each fault cause according to a correspondence between fault causes and fault features and a fault feature of a transmission line.

[0103] In this example, for a transmission line fault that has occurred, the fault recording data and meteorological data before and after the transmission line fault is obtained, for example, at least two fault characteristics are obtained to construct a fault characteristic sample set, and the fault cause corresponding to each fault characteristic or each group of fault characteristics in the fault sample set is obtained. As shown in Table 2, the fault characteristics include weather, time period, season, transition resistance value, transition resistance volt-ampere characteristic, zero-sequence current harmonic content, zero-sequence current DC content, and reclosing conditions, etc.; the fault causes include lightning strikes, wildfires, wind deviation, and tree obstacles, etc. The fault sample set and the fault cause corresponding to each fault characteristic or each group of fault characteristics in the fault sample set are statistically analyzed to obtain the corresponding relationship between the fault cause and the fault characteristic.

[0104] In this embodiment, after obtaining multiple fault features of the transmission line, the fault features can be grouped according to the correspondence between the fault cause and the fault feature to obtain multiple fault feature groups, each of which includes a fault feature corresponding to the fault cause.

[0105] S302: Input the target fault feature corresponding to each fault cause into the weighted naive Bayes model to obtain the classification error rate of each fault cause.

[0106] In this embodiment, there is a one-to-one correspondence between the fault cause and the target fault feature. The fault feature value and the fault cause are input into the weighted naive Bayes model to obtain the classification error rate of each fault cause. For example, a total of 8 fault features are obtained, fault causes Y1, Y2, Y3 and Y4. Each fault cause and the corresponding target fault feature are respectively input into the weighted naive Bayes model to obtain the classification error rates of fault causes Y1, Y2, Y3 and Y4.

[0107] S303: Determine the fault cause corresponding to the minimum classification error rate as the target fault cause of the transmission line.

[0108] In this embodiment, the classification error rates of the fault causes Y1, Y2, Y3 and Y4 are obtained respectively according to S402; the classification error rates of the four fault causes are compared according to their sizes, and the fault cause corresponding to the smallest classification error rate is determined as the target fault cause of the transmission line. For example, the classification error rates of the fault causes Y1, Y2, Y3 and Y4 are 0.6, 0.7, 0.3 and 0.5 respectively, then the fault cause Y3 is the target fault cause.

[0109] In this embodiment, according to the correspondence between the fault cause and the fault feature, and the fault feature of the transmission line, the target fault feature corresponding to each fault cause is determined, each fault cause and the corresponding target fault feature are input into the weighted naive Bayes model to obtain the classification error rate of each fault cause, and the fault cause corresponding to the minimum classification error rate is determined as the target fault cause of the transmission line. The classification error rate of each fault cause can be accurately calculated through the weighted naive Bayes model, and the fault cause corresponding to the minimum classification error rate is determined as the target fault cause of the transmission line, so that the final determined target fault cause is more prepared.

[0110] From the above Figure 2 and Figure 3 In the embodiment, the process of determining the fault cause of the power transmission line in the power grid system by using the weighted naive Bayesian model preset according to the fault characteristics is mainly introduced. The following mainly introduces the training method of the weighted naive Bayesian model. Figure 4 As shown, the training method of the weighted naive Bayes model includes:

[0111] S401, obtaining a fault feature sample set of a transmission line; the fault feature sample set includes sample fault features and actual fault causes corresponding to the sample fault features.

[0112] In this embodiment, fault recording data and meteorological data within a historical period can be obtained, and the historical fault recording data and meteorological data can be analyzed to obtain each historical fault feature and the corresponding fault cause, thereby obtaining sample fault features and the corresponding actual fault causes as a fault feature sample set.

[0113] S402: training a preset initial naive Bayesian model according to the fault feature sample set based on a composite algorithm to obtain a weighted naive Bayesian model.

[0114] In this embodiment, the initial naive Bayes model is trained using the fault feature sample set to obtain a naive Bayes model, and the weight of each attribute of the naive Bayes model is obtained by training based on the composite algorithm and the fault feature sample set, thereby obtaining a weighted naive Bayes model.

[0115] For example, the initial naive Bayes model is trained using the fault feature sample set to obtain a naive Bayes model; based on the fault features, 9 groups of initial weights are randomly generated based on the composite algorithm, and the classification error rate of the weighted naive Bayes classification algorithm under the initial weights is calculated; the composite algorithm is used for iterative calculation to establish an optimization model with the minimum fault cause classification error rate as the objective function:

[0116]

[0117] Where f(w) represents the fault cause classification error rate, which is equal to the number of samples with classification errors during training divided by the total number of training samples; w is a decision vector whose element value represents the weight of each fault feature w k .

[0118] The optimization model of the objective function is solved with the minimum fault cause classification error rate as the condition, the weight of each fault feature is obtained, and the weight of each fault feature and the minimum classification error rate are output.

[0119] In this embodiment, when optimizing the naive Bayes model, a composite algorithm is used to determine the weights of the fault features in the weighted naive Bayes model, so that the weighted naive Bayes model can achieve a higher classification error rate than the naive Bayes model even if fewer fault feature values ​​are input. Therefore, the weighted naive Bayes model based on the composite algorithm is more stable and has a higher classification effect.

[0120] Above Figure 4 The main introduction is to the training method of the weighted naive Bayes model. The following focuses on the process of training the preset initial naive Bayes model based on the composite algorithm according to the fault feature sample set to obtain the weighted naive Bayes model. Figure 5 As shown, the following steps are included:

[0121] S501, randomly extracting a preset number of fault feature samples from a fault feature sample set as a first training set, and taking the remaining fault feature samples as a second training set.

[0122] In this embodiment, the fault feature sample set can be divided into two training sets, and a preset number of fault sample sets can be randomly selected as the first training set by a random sampling method, and the rest are the second training set. For example, a total of 268 transmission line fault samples with different fault cause types in a regional power grid are collected as samples for analysis, and 168 samples are randomly selected as the first training set by a random sampling method, and the remaining 100 samples are used as the second training set.

[0123] Optionally, the fault feature sample set can also be divided into a training set and a test set, wherein the training set is used to train a weighted naive Bayesian model, and the test set is used to test the performance of the trained weighted naive Bayesian model. Furthermore, the training set can also be divided into a first training set and a second training set. For example, a total of 268 transmission line fault samples with different types of fault causes in a regional power grid are collected as samples for analysis, and 168 samples are randomly selected from them as a training set using a random sampling method, and the remaining 100 samples are used as a test set. Furthermore, 126 samples are randomly selected from the training set as the first training set, and the remaining 42 samples are used as the second training set.

[0124] S502: Train the initial naive Bayes model according to the first training set to obtain a naive Bayes model.

[0125] In this embodiment, the fault features and corresponding fault causes in the first training set are input into the initial naive Bayes model, and the parameters of the initial naive Bayes model are iteratively optimized according to the output classification error rate of the fault cause to obtain the naive Bayes model.

[0126] S503: Train the naive Bayes model according to the second training set based on a composite algorithm to obtain a weighted naive Bayes model.

[0127] In this embodiment, based on the second training set, the classification error rate of the weighted naive Bayes model under the initial weights is first calculated. With the minimum classification error rate as the objective function, the weights of the weighted naive Bayes model are continuously optimized using a composite algorithm, and finally the cause of the fault is determined based on the test set.

[0128] In one embodiment, the remaining 42 fault samples are used as the second training set and input into the weighted naive Bayes model, and the classification error rate of the weighted naive Bayes classification algorithm under the initial weights is first calculated. The classification error rates of the 42 fault samples are shown in Table 3:

[0129] Table 3 Comparison of classification error rates of training samples

[0130] algorithm CA_WNB WNB NB Classification error rate 8.7％ 9.4％ 9.7％

[0131] Then, the composite algorithm is used for iterative calculation to optimize the weighted naive Bayesian network model for training the weights of fault features. The weights of each fault feature after optimization by the composite algorithm are shown in Table 4:

[0132] Table 4. Eigenvalues ​​of each fault weight after optimization

[0133]

[0134] Furthermore, the weighted naive Bayes model is obtained by using the optimized weights of each fault feature. The weighted naive Bayes model is tested using a test set consisting of 100 test samples. The test results are shown in Table 5:

[0135] Table 5 Comparison of classification error rates of test samples

[0136] algorithm CA_WNB WNB NB Classification error rate 8.5％ 9.5％ 9.8％

[0137] For example, in actual working conditions, at 6:58 on June 21, 2021, phase B of a 500kV line tripped. According to the fault characteristic value calculated after the transmission line fault, the weighted naive Bayes based on the composite algorithm was used to predict the fault cause. The prediction results are as follows:

[0138] Table 6 Fault characteristic values ​​and fault cause prediction results corresponding to the tripping of a 500 kV line

[0139]

[0140] After the accident, the transmission line management unit organized operation and maintenance personnel to investigate the accident. Based on the comparison of lightning strike data from the lightning location system, traveling wave ranging location results, and on-site fault traces, it was determined that the fault was caused by a lightning strike, which was consistent with the prediction results of the method of this application.

[0141] It can be seen from Tables 3 and 5 that, for training samples and test samples, after adopting the CA_WNB algorithm, the classification error rate drops from 9.4% and 9.5% (classification results of the WNB algorithm) to 8.7% and 8.5% respectively. Regardless of training samples or test samples, the classification error rate of the CA_WNB algorithm is lower than that of the WNB algorithm and the NB algorithm. It can be seen from Table 6 that according to the actual situation, after the accident occurs, the manual investigation results are consistent with the prediction results of the method of the present application, thereby verifying the correctness and effectiveness of the algorithm of the present application, and it has certain practical engineering application value.

[0142] Above Figure 5 This paper mainly introduces the process of training the preset initial naive Bayes model based on the fault feature sample set based on the composite algorithm to obtain the weighted naive Bayes model. Next, the paper focuses on the process of training the naive Bayes model based on the second training set based on the composite algorithm to obtain the weighted naive Bayes model. Figure 6 As shown, the following steps are included:

[0143] S601: construct an initial composite shape according to sample fault features in a second training set.

[0144] In this example, q vertices satisfying the constraint conditions are given or randomly generated to form an initial composite shape, where n+1≤q≤2n, and n is the number of independent variables, which is determined according to the number of fault features selected in actual situations. For example, this embodiment selects 8 fault features, namely, weather, time period, season, transition resistance value, transition resistance volt-ampere characteristic, zero-sequence current DC content, zero-sequence current harmonic content, and reclosing condition, so n=8.

[0145] S602, determining the weight of the best point and the weight of the worst point according to the objective function value of each vertex in the initial composite shape.

[0146] According to the q randomly generated vertices that meet the constraints, the objective function value of each vertex of the composite shape is calculated, that is, the classification error rate of the training samples under each group of weights, and their sizes are compared to find the best point w respectively. L , the worst point w H ,Right now

[0147]

[0148] According to the number of fault features and the q constraint condition, the number of vertices of the composite algorithm is selected as q = 9. For each fault feature, 9 groups of initial weights are randomly generated. The classification error rate of the training samples under each group of initial weights is calculated and compared. The best point w is obtained. L , the worst point w H .

[0149] S603, calculating the reflection point according to the weight of the worst point and the weight of the center point; the center point is the center point of other vertices except the worst point.

[0150] Use the reflection points to continuously find the optimal weights and calculate the center point w according to the worst point C , center point w C To remove the worst point w H The center point of the (q-1) outer vertices,

[0151]

[0152] In general, the worst point w H and the center point w C The direction of the line is the descending direction of the objective function.

[0153] S604, performing iterative calculations according to the objective function values ​​of the reflection points and the objective function values ​​of the worst points until an iteration termination condition is satisfied, and determining the weight of the best point satisfying the iteration termination condition as the weight of the fault cause in the naive Bayes model to obtain a weighted naive Bayes model.

[0154] The iteration termination condition includes that the root mean square value of the difference between the objective function value of each vertex of the composite shape and the objective function value of the best point is less than a preset accuracy value.

[0155] Optionally, the iteration termination condition is that the root mean square value of the difference between the function value of each vertex and the optimal point is less than the precision value, that is,

[0156]

[0157] For example, if the convergence accuracy ε=10-3 is determined, if the termination condition is not met, then return to step S602 to continue the next iteration; otherwise, the best point w of the composite shape is L and its function value f(w L ) is output as the optimal solution.

[0158] Furthermore, if Figure 7 As shown, the iterative operation mainly includes the following steps:

[0159] S701, determining whether the reflection point is a feasible point.

[0160] In this embodiment, according to the center point w C and the worst point w H Calculate the reflection point w R , when the reflection point w R When certain preset conditions are met, the reflection point is judged to be a feasible point; when the reflection point w R If certain preset conditions are not met, the reflection point is judged as a non-feasible point; the calculation formula for the reflection point is:

[0161] w R =w C +α(w C -w H ) (12)

[0162] Where α is the reflection coefficient, which can generally be taken as 1.3.

[0163] Exemplarily, when the reflection point is represented by a coordinate, the preset condition may be a coordinate range, and when the coordinate of the reflection point is within the coordinate range, the reflection point is determined to be a feasible point, otherwise, the reflection point is determined to be a non-feasible point. When the reflection point is represented by a numerical value, the preset condition may be a numerical range, and when the coordinate of the reflection point is within the numerical range, the reflection point is determined to be a feasible point, otherwise, the reflection point is determined to be a non-feasible point.

[0164] S702, if the reflection point is a feasible point and the objective function value of the reflection point is less than the objective function value of the worst point, the reflection point is used to replace the worst point to obtain a new composite shape, and the step of determining the weight of the best point and the weight of the worst point according to the objective function value of each vertex in the new composite shape is returned to be executed.

[0165] In this embodiment, if the reflection point w R is a feasible point, then compare the reflection point w R and the worst point w H The objective function value corresponding to the two points, if f(w R ) <f(w H ), then use the reflection point w R Replace the worst point w H A new composite shape is formed, and one iteration is completed, and the process returns to step S602.

[0166] S703, if the reflection point is a feasible point and the objective function value of the reflection point is not less than the objective function value of the worst point, move the reflection point closer to the center point and calculate a new reflection point until the objective function value of the new reflection point is less than the objective function value of the worst point.

[0167] In this embodiment, if the reflection point w R is a feasible point, then compare the reflection point w R and the worst point w H The objective function value corresponding to the two points, if f(w R )≥f(w H ), the reflection point is moved closer to the center point according to a certain step length, and the step length can be selected according to the actual situation. In this embodiment, the step length can be selected as 0.5, then w R Towards the center point w C Reduce the distance by half, that is, w RP =0.5(w R +w C ), recalculate the new reflection point, and continue this process until f(w R ) <f(w H ).

[0168] S704, if the reflection point is not a feasible point, the reflection point moves closer to the center point until the new reflection point becomes a feasible point, and then the step of making the reflection point a feasible point is performed.

[0169] In this embodiment, if w R If the reflection point is a non-feasible point, the reflection point is moved closer to the center point according to a certain step length, and the step length can be selected according to the actual situation. In this embodiment, the step length can be selected as 0.5, then the reflection point w R Towards the center point w C Reduce the distance by half, that is, w RP =0.5(w R +w C ) until the reflection point w R Stop and repeat w R The step when it is a feasible point, that is, executing the above step S702 or S703.

[0170] In this example, the main work of the composite algorithm optimization method is to generate the initial composite shape and update the composite shape. Generally, only the reflection points in the feasible domain are used as the basic search strategy. The requirements for the initial point are low and the amount of calculation is not large. It can also effectively reduce the calculation variables and find the optimal solution for the variables more quickly. It is particularly suitable for processing small sample data problems.

[0171] Further, such as Figure 8 As shown, the iterative operation may include the following steps:

[0172] Given the number of vertices q, reflection coefficient a, and convergence accuracy ε, generate q points in the feasible domain to construct the initial composite shape; calculate the function value f(w i );Determine the best point w according to the objective function value of each vertex in the initial composite shape L The weight and the worst point w H If the iteration termination condition is met, the weight w of the best point of the fault feature that meets the iteration termination condition is output. L And the classification error rate f(w L ); if the iteration termination condition is not met, then according to the worst point w H The weight and center point w C The weight calculation reflection point w R , the center point is the center point w among the vertices except the worst point C ; Determine the reflection point w R Is it a feasible point? If the reflection point w R is a feasible point, and the reflection point w R The objective function value is less than the worst point w H The objective function value of the reflection point w R Replace the worst point w H , get a new composite shape, and return to execute according to the objective function value of each vertex in the new composite shape to determine the best point w L The weight and the worst point w H The weight of the step; if the reflection point w R is a feasible point, and the reflection point w R The objective function value is not less than the worst point w H The objective function value of the reflection point w R Towards the center point w C Get closer and calculate the new reflection point w R , until the new reflection point w R The objective function value is less than the worst point w H The objective function value; if the reflection point w R is a non-feasible point, then the reflection point w R Move closer to the center point until the new reflection point wR After it is a feasible point, execute the reflection point w R Steps to feasible points.

[0173] The method for determining the cause of a transmission line fault provided in an embodiment of the present application constructs an initial composite shape according to the sample fault features in a second training set, determines the weight of the best point and the weight of the worst point according to the objective function values ​​of each vertex in the initial composite shape, calculates the reflection point according to the weight of the worst point and the weight of the center point, performs iterative operations according to the objective function values ​​of the reflection point and the objective function value of the worst point until an iteration termination condition is met, determines the weight of the best point that meets the iteration termination condition as the weight of the cause of the fault in a naive Bayes model, and obtains a weighted naive Bayes model. In this embodiment, a composite shape algorithm is used to determine the weight of the cause of the fault in the naive Bayes model, so that the performance of the weighted naive Bayes model is better and the obtained classification error rate is more accurate.

[0174] It should be understood that although Figure 2-7 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 2-7 At least part of the steps may include multiple steps or multiple stages. These steps or stages are not necessarily performed at the same time, but can be performed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed in turn or alternately with other steps or at least part of the steps or stages in other steps.

[0175] In one embodiment, Figure 8 As shown, a device for determining the cause of a transmission line fault is provided, comprising: a first acquisition module 11, an extraction module 12 and a determination module 13, wherein:

[0176] The first acquisition module 11 is used to acquire fault recording data and meteorological data of the power grid system;

[0177] An extraction module 12, used for extracting features from the fault recording data and the meteorological data to obtain fault features of the power transmission line of the power grid system;

[0178] The determination module 13 is used to determine the target fault cause of the transmission line in the power grid system according to the fault characteristics and a preset weighted naive Bayes model; the weighted naive Bayes model is obtained according to a weighted naive Bayes classification algorithm based on a composite algorithm.

[0179] In one embodiment, Fig. 9As shown, the determination module 13 includes:

[0180] A first determining unit 131, configured to determine a target fault feature corresponding to each fault cause according to a correspondence between the fault cause and the fault feature and the fault feature of the transmission line;

[0181] An input unit 132, used to input the target fault feature corresponding to each of the fault causes into the weighted naive Bayes model to obtain a classification error rate of each of the fault causes;

[0182] The second determining unit 133 is configured to determine the fault cause corresponding to the minimum classification error rate as the target fault cause of the power transmission line.

[0183] In one embodiment, the weighted naive Bayes model includes a weight of each fault cause, and the weight of each fault cause is determined by using a composite algorithm.

[0184] In one embodiment, Fig.10 As shown, the device for determining the cause of a fault in a power transmission line further includes:

[0185] The second acquisition module 14 is used to acquire a fault feature sample set of the power transmission line; the fault feature sample set includes sample fault features and actual fault causes corresponding to the sample fault features;

[0186] The training module 15 is used to train the preset initial naive Bayes model according to the fault feature sample set based on a composite algorithm to obtain the weighted naive Bayes model.

[0187] In one embodiment, Fig.11 As shown, the training module 15 includes:

[0188] An extraction unit 151 is used to randomly extract a preset number of fault feature samples from the fault feature sample set as a first training set, and use the remaining fault feature samples as a second training set;

[0189] A first training unit 152, configured to train the initial naive Bayes model according to the first training set to obtain a naive Bayes model;

[0190] The second training unit 153 is used to train the naive Bayes model according to the second training set based on the composite algorithm to obtain the weighted naive Bayes model.

[0191] In one embodiment, the second training unit 153 is specifically used to construct an initial composite shape according to the sample fault features in the second training set; determine the weight of the best point and the weight of the worst point according to the objective function value of each vertex in the initial composite shape; calculate the reflection point according to the weight of the worst point and the weight of the center point; the center point is the center point of other vertices except the worst point; perform iterative operations according to the objective function value of the reflection point and the objective function value of the worst point until the iteration termination condition is met, and determine the weight of the best point that meets the iteration termination condition as the weight of the fault cause in the naive Bayes model to obtain the weighted naive Bayes model.

[0192] In one embodiment, the second training unit 153 is specifically used to determine whether the reflection point is a feasible point; if the reflection point is a feasible point, and the objective function value of the reflection point is less than the objective function value of the worst point, the reflection point is used to replace the worst point to obtain a new composite shape, and the step of determining the weight of the best point and the weight of the worst point according to the objective function value of each vertex in the new composite shape is returned to execute; if the reflection point is a feasible point, and the objective function value of the reflection point is not less than the objective function value of the worst point, the reflection point is moved closer to the center point, and a new reflection point is calculated until the objective function value of the new reflection point is less than the objective function value of the worst point; if the reflection point is not a feasible point, the reflection point is moved closer to the center point until the new reflection point is a feasible point, and then the step of making the reflection point a feasible point is executed.

[0193] In one embodiment, the iteration termination condition includes that the root mean square value of the difference between the objective function value of each vertex of the composite shape and the objective function value of the best point is less than a preset accuracy value.

[0194] The implementation principle and beneficial effects of the device for determining the cause of a fault in a power transmission line provided in the embodiment of the present application can refer to the implementation principle and beneficial effects of the embodiment of the method for determining the cause of a fault in a power transmission line described above, and will not be repeated here.

[0195] The specific definition of the device for determining the cause of a fault in a power transmission line can be found in the definition of the method for determining the cause of a fault in a power transmission line mentioned above, and will not be repeated here. Each module in the device for determining the cause of a fault in a power transmission line can be implemented in whole or in part by software, hardware, or a combination thereof. Each of the modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each of the modules above.

[0196] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Fig.12As shown. The computer device includes a processor, a memory and a network interface connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data such as fault recording data, meteorological data, fault characteristics, and classification error rate. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for determining the cause of a transmission line fault is implemented.

[0197] Those skilled in the art will understand that Fig.12 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0198] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:

[0199] Obtain fault recording data and meteorological data of the power grid system;

[0200] Extracting features from the fault recording data and the meteorological data to obtain fault features of the power transmission line of the power grid system;

[0201] According to the fault characteristics and a preset weighted naive Bayesian model, a target fault cause of a transmission line in the power grid system is determined; the weighted naive Bayesian model is obtained according to a weighted naive Bayesian classification algorithm based on a composite algorithm.

[0202] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0203] The weighted naive Bayes model preset according to the fault characteristics determines the fault cause of the power transmission line in the power grid system, including:

[0204] Determine a target fault feature corresponding to each fault cause according to the correspondence between the fault cause and the fault feature and the fault feature of the transmission line;

[0205] Inputting the target fault feature corresponding to each of the fault causes into the weighted naive Bayes model to obtain the classification error rate of each of the fault causes;

[0206] The fault cause corresponding to the minimum classification error rate is determined as the target fault cause of the transmission line.

[0207] In one embodiment, the weighted naive Bayes model includes a weight of each fault cause, and the weight of each fault cause is determined by using a composite algorithm.

[0208] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0209] Acquire a fault feature sample set of a transmission line; the fault feature sample set includes sample fault features and actual fault causes corresponding to the sample fault features;

[0210] The preset initial naive Bayesian model is trained according to the fault feature sample set based on a composite algorithm to obtain the weighted naive Bayesian model.

[0211] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0212] Randomly extracting a preset number of fault feature samples from the fault feature sample set as a first training set, and using the remaining fault feature samples as a second training set;

[0213] Training the initial naive Bayes model according to the first training set to obtain a naive Bayes model;

[0214] The naive Bayes model is trained according to the second training set based on the composite algorithm to obtain the weighted naive Bayes model.

[0215] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0216] Constructing an initial composite shape according to the sample fault features in the second training set;

[0217] Determining the weight of the best point and the weight of the worst point according to the objective function value of each vertex in the initial composite shape;

[0218] Calculate the reflection point according to the weight of the worst point and the weight of the center point; the center point is the center point of other vertices except the worst point;

[0219] According to the objective function value of the reflection point and the objective function value of the worst point, an iterative operation is performed until an iteration termination condition is met, and the weight of the best point that meets the iteration termination condition is determined as the weight of the fault cause in the naive Bayes model to obtain the weighted naive Bayes model.

[0220] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0221] Determining whether the reflection point is a feasible point;

[0222] If the reflection point is a feasible point and the objective function value of the reflection point is less than the objective function value of the worst point, the reflection point is used to replace the worst point to obtain a new composite shape, and the step of determining the weight of the best point and the weight of the worst point according to the objective function value of each vertex in the new composite shape is returned;

[0223] If the reflection point is a feasible point and the objective function value of the reflection point is not less than the objective function value of the worst point, the reflection point is moved closer to the center point, and a new reflection point is calculated until the objective function value of the new reflection point is less than the objective function value of the worst point;

[0224] If the reflection point is not a feasible point, the reflection point moves closer to the center point until the new reflection point becomes a feasible point, and then the step of making the reflection point a feasible point is performed.

[0225] In one embodiment, the iteration termination condition includes that the root mean square value of the difference between the objective function value of each vertex of the composite shape and the objective function value of the best point is less than a preset accuracy value.

[0226] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0227] Obtain fault recording data and meteorological data of the power grid system;

[0228] Extracting features from the fault recording data and the meteorological data to obtain fault features of the power transmission line of the power grid system;

[0229] According to the fault characteristics and a preset weighted naive Bayesian model, a target fault cause of a transmission line in the power grid system is determined; the weighted naive Bayesian model is obtained according to a weighted naive Bayesian classification algorithm based on a composite algorithm.

[0230] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0231] The weighted naive Bayes model preset according to the fault characteristics determines the fault cause of the power transmission line in the power grid system, including:

[0232] Determine a target fault feature corresponding to each fault cause according to the correspondence between the fault cause and the fault feature and the fault feature of the transmission line;

[0233] Inputting the target fault feature corresponding to each of the fault causes into the weighted naive Bayes model to obtain the classification error rate of each of the fault causes;

[0234] The fault cause corresponding to the minimum classification error rate is determined as the target fault cause of the transmission line.

[0235] In one embodiment, the weighted naive Bayes model includes a weight of each fault cause, and the weight of each fault cause is determined by using a composite algorithm.

[0236] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0237] Acquire a fault feature sample set of a transmission line; the fault feature sample set includes sample fault features and actual fault causes corresponding to the sample fault features;

[0238] The preset initial naive Bayesian model is trained according to the fault feature sample set based on a composite algorithm to obtain the weighted naive Bayesian model.

[0239] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0240] Randomly extracting a preset number of fault feature samples from the fault feature sample set as a first training set, and using the remaining fault feature samples as a second training set;

[0241] Training the initial naive Bayes model according to the first training set to obtain a naive Bayes model;

[0242] The naive Bayes model is trained according to the second training set based on the composite algorithm to obtain the weighted naive Bayes model.

[0243] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0244] Constructing an initial composite shape according to the sample fault features in the second training set;

[0245] Determining the weight of the best point and the weight of the worst point according to the objective function value of each vertex in the initial composite shape;

[0246] Calculate the reflection point according to the weight of the worst point and the weight of the center point; the center point is the center point of other vertices except the worst point;

[0247] According to the objective function value of the reflection point and the objective function value of the worst point, an iterative operation is performed until an iteration termination condition is met, and the weight of the best point that meets the iteration termination condition is determined as the weight of the fault cause in the naive Bayes model to obtain the weighted naive Bayes model.

[0248] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0249] Determining whether the reflection point is a feasible point;

[0250] If the reflection point is a feasible point and the objective function value of the reflection point is less than the objective function value of the worst point, the reflection point is used to replace the worst point to obtain a new composite shape, and the step of determining the weight of the best point and the weight of the worst point according to the objective function value of each vertex in the new composite shape is returned;

[0251] If the reflection point is a feasible point and the objective function value of the reflection point is not less than the objective function value of the worst point, the reflection point is moved closer to the center point, and a new reflection point is calculated until the objective function value of the new reflection point is less than the objective function value of the worst point;

[0252] If the reflection point is not a feasible point, the reflection point moves closer to the center point until the new reflection point becomes a feasible point, and then the step of making the reflection point a feasible point is performed.

[0253] In one embodiment, the iteration termination condition includes that the root mean square value of the difference between the objective function value of each vertex of the composite shape and the objective function value of the best point is less than a preset accuracy value.

[0254] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0255] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0256] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A method for determining the cause of a transmission line fault, characterized in that: The method comprises: Obtain fault recording data and meteorological data of the power grid system; Extracting features from the fault recording data and the meteorological data to obtain fault features of the power transmission line of the power grid system; Determining target fault causes of transmission lines in the power grid system according to the fault characteristics and a preset weighted naive Bayes model; The training method of the weighted naive Bayes model includes: Acquire a fault feature sample set of a transmission line; the fault feature sample set includes sample fault features and actual fault causes corresponding to the sample fault features; A preset number of fault feature samples are randomly selected from the fault feature sample set as the first training set, and the remaining fault feature samples are used as the second training set; the preset initial naive Bayes model is trained according to the first training set to obtain the naive Bayes model; an initial composite shape is constructed according to the sample fault features in the second training set; the weight of the best point and the weight of the worst point are determined according to the objective function value of each vertex in the initial composite shape; the reflection point is calculated according to the weight of the worst point and the weight of the center point; the center point is the center point of other vertices except the worst point; according to the objective function value of the reflection point and the objective function value of the worst point, an iterative operation is performed until the iteration termination condition is met, and the weight of the best point that meets the iteration termination condition is determined as the weight of the fault cause in the naive Bayes model to obtain the weighted naive Bayes model.

2. The method according to claim 1, characterized in that The weighted naive Bayes model preset according to the fault characteristics determines the fault cause of the power transmission line in the power grid system, including: Determine a target fault feature corresponding to each fault cause according to the correspondence between the fault cause and the fault feature and the fault feature of the transmission line; Inputting the target fault feature corresponding to each of the fault causes into the weighted naive Bayes model to obtain the classification error rate of each of the fault causes; The fault cause corresponding to the minimum classification error rate is determined as the target fault cause of the transmission line.

3. The method according to claim 1 or 2, characterized in that: The weighted naive Bayes model includes a weight of each fault cause, and the weight of each fault cause is determined by using a composite algorithm.

4. The method according to claim 1, characterized in that The iterative operation is performed according to the objective function value of the reflection point and the objective function value of the worst point, comprising: Determining whether the reflection point is a feasible point; If the reflection point is a feasible point and the objective function value of the reflection point is less than the objective function value of the worst point, the reflection point is used to replace the worst point to obtain a new composite shape, and the step of determining the weight of the best point and the weight of the worst point according to the objective function value of each vertex in the new composite shape is returned; If the reflection point is a feasible point and the objective function value of the reflection point is not less than the objective function value of the worst point, the reflection point is moved closer to the center point, and a new reflection point is calculated until the objective function value of the new reflection point is less than the objective function value of the worst point; If the reflection point is not a feasible point, the reflection point moves closer to the center point until the new reflection point becomes a feasible point, and then the step of making the reflection point a feasible point is performed.

5. The method according to claim 1, characterized in that The iteration termination condition includes that the root mean square value of the difference between the objective function value of each vertex of the composite shape and the objective function value of the best point is less than a preset accuracy value.

6. A device for determining the cause of a transmission line fault, characterized in that: The device comprises: The first acquisition module is used to acquire fault recording data and meteorological data of the power grid system; An extraction module, used for extracting features from the fault recording data and the meteorological data to obtain fault features of the power transmission line of the power grid system; A determination module, used to determine the target fault cause of the transmission line in the power grid system according to the fault characteristics and a preset weighted naive Bayes model; A second acquisition module is used to acquire a fault feature sample set of the power transmission line; the fault feature sample set includes sample fault features and actual fault causes corresponding to the sample fault features; A training module is used to randomly extract a preset number of fault feature samples from the fault feature sample set as the first training set, and use the remaining fault feature samples as the second training set; train the preset initial naive Bayes model according to the first training set to obtain a naive Bayes model; construct an initial composite shape according to the sample fault features in the second training set; determine the weight of the best point and the weight of the worst point according to the objective function value of each vertex in the initial composite shape; calculate the reflection point according to the weight of the worst point and the weight of the center point; the center point is the center point of other vertices except the worst point; perform iterative calculations according to the objective function value of the reflection point and the objective function value of the worst point until the iteration termination condition is met, and determine the weight of the best point that meets the iteration termination condition as the weight of the fault cause in the naive Bayes model to obtain the weighted naive Bayes model.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 5 are implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images