Power transmission line overhead line state determination method, device, equipment, medium and program product

Abstract

The invention relates to a state determination method and device for an overhead line of a power transmission line, computer equipment, a computer readable storage medium and a computer program product. The method comprises the steps of obtaining to-be-detected data sent by at least one edge node, wherein each piece of to-be-detected data comprises at least one piece of state data; data features of each piece of state data are determined, a reference data feature set corresponding to the overhead line of the power transmission line is obtained, and the reference data feature set comprises reference data features corresponding to each data type; for each piece of state data, querying a target data type matched with the state data according to the difference between the data feature of the state data and each reference data feature; and calling the state prediction model based on the multiple pieces of state data and the target data types of the multiple pieces of state data, and determining the operation state of the power transmission line overhead line, thereby improving the operation state prediction accuracy.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method, device, computer equipment, computer-readable storage medium and computer program product for determining the state of an overhead line of a power transmission line. Background Art

[0002] With the development of high voltage technology and the gradual expansion of the power grid, the total length of transmission lines is also increasing rapidly. Since transmission lines, especially overhead high-voltage transmission lines, are always exposed to the air, extreme weather events such as heavy snow, strong winds and heavy fog can cause great damage to them. Therefore, it is necessary to perform status detection on overhead transmission lines.

[0003] In the related technology, multiple types (ice tension, temperature or fault location, etc.) and multiple sets of online detection terminals are deployed on the transmission line to sense the operating status of the equipment. Then, a single set of online terminals with points as detection units are used to detect the overhead lines of the transmission lines. However, the detection dimension of a single set of online terminals with points as monitoring units is single, and it is difficult to accurately predict the status of the overhead lines of the transmission lines. Summary of the invention

[0004] Based on this, it is necessary to provide a method, device, computer equipment, computer-readable storage medium and computer program product for determining the state of overhead lines of transmission lines, which can improve the accuracy of prediction of the state of overhead lines of transmission lines in response to the above technical problems.

[0005] In a first aspect, the present application provides a method for determining the state of an overhead line of a power transmission line, comprising:

[0006] Acquire data to be detected sent by at least one edge node, each data to be detected is obtained by collecting data from a preset position of an overhead line of a power transmission line by a collection device corresponding to the edge node, and each data to be detected includes at least one state data;

[0007] Determine data features of each state data, and obtain a reference data feature set corresponding to the overhead line of the transmission line, wherein the reference data feature set includes reference data features corresponding to each data type;

[0008] For each state data, according to the difference between the data feature of the state data and each reference data feature, query the target data type matched by the state data;

[0009] Based on a plurality of state data and respective target data types of the plurality of state data, a state prediction model is called to determine the operating state of the overhead line of the transmission line.

[0010] In a second aspect, the present application further provides a device for determining the state of an overhead line of a power transmission line, comprising:

[0011] A first acquisition module is used to acquire data to be detected sent by at least one edge node, each data to be detected is obtained by collecting data from a preset position of an overhead line of a power transmission line by a collection device corresponding to the edge node, and each data to be detected includes at least one state data;

[0012] A second acquisition module is used to determine the data features of each state data, and acquire a reference data feature set corresponding to the overhead line of the transmission line, wherein the reference data feature set includes reference data features corresponding to each data type;

[0013] A query module, configured to query, for each state data, the target data type that matches the state data according to the difference between the data characteristics of the state data and the characteristics of each reference data;

[0014] The determination module is used to call the state prediction model based on multiple state data and the target data types of the multiple state data to determine the operating state of the overhead line of the transmission line.

[0015] In a third aspect, the present application further provides a computer device, including 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:

[0016] Acquire data to be detected sent by at least one edge node, each data to be detected is obtained by collecting data from a preset position of an overhead line of a power transmission line by a collection device corresponding to the edge node, and each data to be detected includes at least one state data;

[0017] Determine data features of each state data, and obtain a reference data feature set corresponding to the overhead line of the transmission line, wherein the reference data feature set includes reference data features corresponding to each data type;

[0018] For each state data, according to the difference between the data feature of the state data and each reference data feature, query the target data type matched by the state data;

[0019] Based on a plurality of state data and respective target data types of the plurality of state data, a state prediction model is called to determine the operating state of the overhead line of the transmission line.

[0020] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:

[0021] Acquire data to be detected sent by at least one edge node, each data to be detected is obtained by collecting data from a preset position of an overhead line of a power transmission line by a collection device corresponding to the edge node, and each data to be detected includes at least one state data;

[0022] Determine data features of each state data, and obtain a reference data feature set corresponding to the overhead line of the transmission line, wherein the reference data feature set includes reference data features corresponding to each data type;

[0023] For each state data, according to the difference between the data feature of the state data and each reference data feature, query the target data type matched by the state data;

[0024] Based on a plurality of state data and respective target data types of the plurality of state data, a state prediction model is called to determine the operating state of the overhead line of the transmission line.

[0025] In a fifth aspect, the present application further provides a computer program product, including a computer program, which implements the following steps when executed by a processor:

[0026] Acquire data to be detected sent by at least one edge node, each data to be detected is obtained by collecting data from a preset position of an overhead line of a power transmission line by a collection device corresponding to the edge node, and each data to be detected includes at least one state data;

[0027] Determine data features of each state data, and obtain a reference data feature set corresponding to the overhead line of the transmission line, wherein the reference data feature set includes reference data features corresponding to each data type;

[0028] For each state data, according to the difference between the data feature of the state data and each reference data feature, query the target data type matched by the state data;

[0029] Based on a plurality of state data and respective target data types of the plurality of state data, a state prediction model is called to determine the operating state of the overhead line of the transmission line.

[0030] The above-mentioned method, device, computer equipment, computer-readable storage medium and computer program product for determining the state of the overhead line of the transmission line obtain the data to be detected sent by at least one edge node, each of which is obtained by collecting data from a preset position of the overhead line of the transmission line by the collection device of the corresponding edge node, and each of which includes at least one state data. That is to say, according to the collection devices of one or more edge nodes, it is possible to complete comprehensive data collection of the overhead line deployed in the transmission line to ensure the accuracy of the state estimation. Determine the data features of each state data, and obtain a reference data feature set corresponding to the overhead line of the transmission line, the reference data feature set includes reference data features corresponding to each data type; for each state data, query the target data type matched by the state data according to the difference between the data features of the state data and each reference data feature. Thus, the data type of each state data is quickly and accurately identified, so that based on multiple state data and the target data types of each of the multiple state data, the state prediction model is called, and the operation state of the overhead line of the transmission line can be comprehensively predicted from multiple dimensions with the help of multi-dimensional detail information, thereby ensuring the accuracy of the state prediction of the overhead line of the transmission line. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0032] Figure 1 A diagram of an application environment of a method for determining a state of an overhead line of a power transmission line in one embodiment;

[0033] Figure 2 A schematic flow chart of a method for determining a state of an overhead line of a power transmission line in one embodiment;

[0034] Figure 3 A schematic diagram of a flow chart of a step of determining a target data type in one embodiment;

[0035] Figure 4 It is a structural block diagram of a device for determining the state of an overhead line of a power transmission line in one embodiment;

[0036] Figure 5 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0037] 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.

[0038] The method for determining the state of an overhead line of a power transmission line provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. Each edge node in the edge node cluster 102 communicates with the target server 104, wherein the edge node cluster is used to process data of the overhead line of the transmission line, each edge node can be regarded as an edge server, and each edge node has a corresponding collection device, which is used to collect data from the preset position of the overhead line of the transmission line. The target server 104 is used to predict the state of the overhead line of the transmission line.

[0039] In one embodiment, the target server 104 obtains data to be detected sent by at least one edge node, each data to be detected is obtained by collecting data from a preset position of an overhead line of a transmission line by a collection device corresponding to the edge node, and each data to be detected includes at least one state data; the target server 104 determines data features of each state data, and obtains a reference data feature set corresponding to the overhead line of the transmission line, the reference data feature set includes reference data features corresponding to each data type; for each state data, the target server 104 queries the target data type that matches the state data based on the difference between the data features of the state data and each reference data feature; the target server 104 calls a state prediction model based on multiple state data and the respective target data types of the multiple state data to determine the operating state of the overhead line of the transmission line.

[0040] Among them, the target server 104 and the edge node are different servers. The server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides cloud computing services.

[0041] In an exemplary embodiment, Figure 2 As shown, a method for determining the state of an overhead line of a transmission line is provided, and the method is applied to Figure 1 Taking the target server 104 in the example as an example, the method includes the following steps 202 to 208. Among them:

[0042] Step 202, obtaining data to be detected sent by at least one edge node, each data to be detected is obtained by collecting data from a preset position of an overhead line of a transmission line by a collection device corresponding to the edge node, and each data to be detected includes at least one state data.

[0043] Corresponding acquisition devices are deployed at multiple locations of the overhead line of the transmission line, and each acquisition device is connected to a corresponding edge node. The acquisition device can be a camera, a current acquisition device, a voltage acquisition device, etc. The data to be detected includes at least one state data, and each state data has a different data type. For example, the state data can be image type data, that is, it can be an image, or it can be numerical type data, such as current, voltage, temperature, humidity, etc.

[0044] Optionally, for each edge node, the edge node obtains status data sent by at least one acquisition device connected thereto, and the edge node sends the at least one status data to the target server.

[0045] It should be noted that the edge node is connected to the acquisition device, and each acquisition device is used to collect data of the corresponding data type. Therefore, the edge node knows the data type of the acquired state data. Since the target server communicates with the edge node, the target server does not know what data type the state data is. In addition, in the process of data transmission, the transmission form can be binary transmission, such as obtaining the gray value of the image and converting it into a binary signal. Similarly, converting the current or voltage and other numerical values ​​into binary signals, then the target server cannot know what kind of data the state is after obtaining any state data, whether it reflects image information or current information, etc. Therefore, after obtaining multiple state data, the target server needs to identify the data type of the state data in order to make reasonable predictions according to the matching prediction method.

[0046] Step 204, determine the data features of each state data, and obtain a reference data feature set corresponding to the overhead line of the transmission line, the reference data feature set including reference data features corresponding to each data type.

[0047] Among them, the data features of the status data reflect the detailed information. For example, if the status data is an image, the corresponding data features include the appearance information of the overhead line of the transmission line to reflect the offset, connection stability, etc. of the overhead line of the transmission line. For another example, if the status data is a current value of a numerical type, the corresponding data features include the current situation of the overhead line of the transmission line to reflect whether the current of the overhead line of the transmission line is abnormal. For this reason, it can be understood that the data features of status data of different data types are different. The reference data feature set is a pre-set feature set about the overhead line of the transmission line.

[0048] The reference data feature set includes reference data features of multiple data types related to overhead lines of power transmission lines. In one embodiment, the reference data feature determination step of each data type is as follows: for each data type, multiple reference data belonging to the data type are obtained, and the data features of each reference data are extracted respectively using a feature extraction model, and the data features of the multiple reference data are integrated to obtain the reference data features of the data type. For example, the mean of the data features of the multiple reference data is taken as the reference data feature of the data type. Among them, the feature extraction model is used for feature extraction, and is a trained neural network model, or it can be a data statistical model. For example, for image types, the data statistical model counts the memory size of the image. For another example, for numerical types, such as current, the data statistical model counts the current value, that is, the statistical data value size.

[0049] In one embodiment, for each state data, the target server uses a data feature extraction model to extract features from the state data to obtain corresponding data features.

[0050] Step 206 : for each state data, according to the difference between the data feature of the state data and the feature of each reference data, the target data type that matches the state data is searched.

[0051] Among them, the smaller the difference is, the more similar the data feature is to the corresponding reference data feature, and the greater the probability that the state data belongs to the data type of the reference data feature.

[0052] In one embodiment, for each state data, the target data type that matches the state data is queried based on the differences between the data features of the state data and the features of each reference data, including: for each state data, the differences between the data features of the state data and the features of each reference data are calculated, and the target data type that matches the state data is queried with the goal of minimizing the differences.

[0053] Among them, the difference between the data characteristics of the state data and the reference data characteristics of the target data type is the smallest.

[0054] Exemplarily, for each state data, the target server calculates the vector product between the data feature of the state data and each reference data feature, and queries the target data type that matches the state data with the goal of minimizing the vector product. In this example, both the data feature and the reference data feature can be regarded as vectors. For this reason, the smallest vector product means that the more similar the data feature is to the corresponding reference data feature, that is, the smaller the difference is, the greater the probability of the data type belonging to the reference data feature is.

[0055] In this embodiment, by minimizing the difference as the goal, the reference data feature with the smallest difference from the data feature of the status data can be accurately queried, and the data type of the status data can be accurately classified as belonging to the data type of the reference data feature, thereby improving the accuracy of data type identification of the status data.

[0056] In one embodiment, Figure 3 As shown, it is a flowchart of the target data type determination step in one embodiment. For each state data, the difference between the data feature of the state data and each reference data feature is calculated, and the target data type matching the state data is queried with the goal of minimizing the difference, including:

[0057] Step 302: Based on multiple data types involved in the reference data feature set, multiple groups of different candidate combinations are determined, and each group of candidate combinations defines the data type of each state data.

[0058] Optionally, the target server determines multiple data types involved, and freely combines multiple data types and each state data to obtain multiple groups of different candidate combinations. Exemplarily, the data type includes m data types, and there are n state data. For this reason, each state data can have m choices, and any one of the data types L1 to L2 can be selected. Each candidate combination includes the data type of each of the n state data. For example, there are 3 state data, namely data1 to data3, and a candidate combination: L1, L2, L3, indicating that the data types of data1 to data3 are L1, L2, L3 respectively; or, the candidate combination: L1, L1, L3, indicating that the data types of data1 to data3 are L1, L1, L3 respectively, and so on, and there are multiple candidate combinations.

[0059] Step 304, for each group of candidate combinations, based on the data type defined for each state data in the candidate combination, calculate the difference between the data feature of the state data and the reference data feature corresponding to the data type to obtain the difference corresponding to the state data, and superimpose the differences corresponding to the various state data to obtain the total difference corresponding to the candidate combination.

[0060] Exemplarily, for each group of candidate combinations, according to the data type defined for each state data in the candidate combination, the difference between the data feature of the state data and the reference data feature corresponding to the data type is calculated, and the absolute value of the difference is determined as the difference corresponding to the state data, and the differences corresponding to each state data are superimposed to obtain the total difference corresponding to the candidate combination. Among them, the absolute value can more clearly show the degree of difference between the data feature and each reference data feature.

[0061] Step 306: taking the candidate combination with the smallest total difference as the target combination, and determining the target data type of each state data based on the target combination.

[0062] For example, the target server pre-establishes a classification function f(n), and determines the target data type of each state data based on the classification function:

[0063]

[0064] Among them, argmin (.) is used to find the minimum parameter. is the data feature of the i-th state data in the n-th candidate combination, The parameter data feature corresponding to the data type of the i-th state data when it is the n-th candidate combination. It should be noted that, The same is true for each candidate combination, and refers to the data features of the i-th state data. m is the number of state data.

[0065] In this embodiment, by traversing the total difference of each candidate combination in turn, the candidate combination with the smallest total difference can be queried to determine the optimal data type classification set, and the target data type matched by each state data is accurately determined.

[0066] Step 208: Based on the plurality of state data and the target data types of the plurality of state data, a state prediction model is called to determine the operating state of the overhead line of the transmission line.

[0067] Among them, the state prediction model is a model built based on a neural network. In one embodiment, the state prediction model can be a neural network model built based on federated learning. Exemplarily, each edge node deploys at least one sub-prediction model, and each sub-prediction model is used to predict the state data of the corresponding data type. The target server obtains the model parameters of each sub-prediction model sent by each edge node to build a state prediction model. It should be noted that the state prediction model is used to make an overall evaluation and prediction of the state of the entire transmission line overhead line to verify whether there are any abnormal parts. It should be noted that since the sub-prediction model can only predict the preset part corresponding to a single edge node, that is, it can only predict part of the transmission line overhead line, and cannot predict all parts. For this reason, it is necessary to build a state prediction model based on the target server that can predict the entire transmission line overhead line.

[0068] In one embodiment, the state prediction model includes a classification network and prediction networks corresponding to each data type. Based on multiple state data and the target data types of each of the multiple state data, the state prediction model is called to determine the operating state of the overhead line of the transmission line, including: inputting each state data and the target data type of each state data into the state prediction model, and based on the target data type corresponding to each state data, determining the prediction network corresponding to each state data through the classification network; for each state data, performing state prediction on the state data through the corresponding prediction network to obtain a corresponding prediction result; and determining the operating state of the overhead line of the transmission line through each prediction result.

[0069] Exemplarily, the target server obtains each state data and the target data type of each state data, and inputs them into the state prediction model. Through the classification network in the state prediction model, the prediction network corresponding to each state data is determined according to each target data type. For each state data, the state data is predicted by the corresponding prediction network to obtain the corresponding prediction result. When each prediction result represents normal, the operation state of the overhead line of the transmission line is determined to be normal. When at least one prediction result represents abnormality, the operation state of the overhead line of the transmission line is determined to be abnormal.

[0070] In one embodiment, when at least one prediction result characterizes an abnormality, the method further includes: determining an edge node from which status data corresponding to at least one prediction result originates, and using a preset part corresponding to the determined edge node as a suspected abnormal part, and generating an alarm information based on the suspected abnormal part to instruct operation and maintenance personnel to troubleshoot the suspected abnormal part.

[0071] In this embodiment, targeted prediction is performed through a prediction network that can match the target data type of each state data, thereby ensuring the accuracy of the state prediction of the overhead line of the transmission line.

[0072] In one embodiment, the step of determining the state prediction model includes: obtaining each model parameter sent by different edge nodes respectively through a model provider, each model parameter is obtained by the corresponding edge node performing model training on the corresponding sub-prediction model according to the training sample of the corresponding data type; for each data type, obtaining at least one model parameter corresponding to the data type through the model provider, fusing at least one model parameter to obtain the corresponding fused model parameter, and constructing a prediction network corresponding to the data type based on the fused model parameter; obtaining the state prediction model through the model provider based on each constructed prediction network.

[0073] Among them, the model provider can be understood as a model providing server, which can be the same as the target server, or a server different from the target server and the edge node. Exemplarily, for each edge node, there is at least one sub-prediction model of a data type. For each data type, through the edge node, according to the training sample of the data type, the sub-prediction model corresponding to the data type is used for model training to obtain a prediction result. According to the difference between the prediction result and the label of the training sample, the model parameters of the sub-prediction model are updated to obtain a trained sub-prediction model, and the model parameters of the trained sub-prediction model are obtained. The edge node sends the model parameters of at least one deployed sub-prediction model to the model provider. After obtaining the model parameters sent by each edge node, the model provider counts the data types involved in the obtained model parameters. For each data type, the model provider performs weighted summation or mean calculation on the model parameters corresponding to the data type to obtain the fusion model parameters of the data type, and constructs the prediction network corresponding to the data type according to the fusion model parameters. Based on the prediction models corresponding to each data type, a state prediction model is constructed.

[0074] For example: the data types involved in the model parameters obtained by the model provider statistics include data type L1. For data type L1, model parameters W1-Wy are selected from the obtained model parameters to correspond to L1. W1 to Wy are derived from edge nodes 1 to edge nodes y, respectively. Therefore, the fusion model parameters of L1 are obtained by weighting W1 to Wy according to the respective weights of edge nodes 1 to edge nodes and summing them:

[0075]

[0076] Among them, W10 is the fusion model parameter of L1, Wx is the model parameter corresponding to the edge node x, x is [1,2], and Nx is the weight of the edge node x.

[0077] In one embodiment, when the data type is an image type, the corresponding sub-prediction model is an image prediction model. For any edge node where the image prediction model is deployed, the training steps of the image prediction model include: constructing an LSTM (Long Short-Term Memory) model as an image prediction model, inputting the historical image sequence into the image prediction model as an image training sample, training the image prediction model, and extracting the hidden state of each time step. , based on the state of the last time step, it is mapped to the category space, and the probability distribution of each category is output through the softmax activation function. The running state corresponding to the maximum probability in the probability distribution is used as the prediction result. According to the difference between the label corresponding to the historical image sequence and the prediction result, the image prediction model is trained to obtain the trained image prediction model. Among them, the hidden state Represents an abstract representation of the characteristics of historical image sequences; the image prediction model is expressed as:

[0078] Forget Gate: ;

[0079] Input Gate: ;

[0080] Output Gate: ;

[0081] in, is the activation function, is the forget gate function, v t is the historical grayscale image data corresponding to time step t, is the hidden state at time step t-1, b f is the bias vector of the forget gate, W f is the weight matrix of the forget gate, is the update gate function, W I is the weight matrix of the update gate, b I is the bias vector of the update gate, is the candidate cell state function, is a compression function that compresses the output value to Between C is the weight matrix of candidate cells, b C is the bias vector of the candidate cell, C t is the cell update function, is the output gate function, W o is the weight function of the output gate, b o is the bias vector of the output gate.

[0082] In one embodiment, when the data type is a numerical type, the corresponding sub-prediction model is a numerical prediction model. For any edge node where a numerical prediction model is deployed, the training steps of the numerical prediction model include:

[0083] Constructing a polynomial regression model as a numerical prediction model , the maximum probability As the dependent variable of the numerical prediction model, the historical numerical dataset corresponding to the image training dataset As an independent variable in numerical prediction models;

[0084] ;

[0085] Where z is the number of data in the historical numerical data set, is the zth data, are the coefficients of the numerical prediction model, is a constant; the numerical prediction model is trained using the historical numerical data set, the coefficients and constants are fitted, and the fitted numerical prediction model is output. It should be noted that after the numerical prediction model is trained, the coefficients of the numerical prediction model are obtained, and the coefficients are used as model parameters to construct the corresponding numerical prediction network (also constructed based on the polynomial regression model).

[0086] In this embodiment, by obtaining the model parameters sent by different edge nodes respectively, the fusion model parameters of different data types are determined to construct a corresponding prediction network. The state prediction model constructed in this way can effectively predict the entire transmission line overhead line and ensure the reliability and effectiveness of the state prediction.

[0087] In one embodiment, when the data type is an image type, the step of determining a training sample of the image type includes: obtaining an initial image set through an edge node corresponding to the image type; filtering out initial images whose grayscale values ​​meet grayscale value conditions from the initial image set, and using the filtered initial images as training samples of the image type.

[0088] As mentioned above, when the data type is an image type, the corresponding sub-prediction model is an image prediction model, and the acquired training samples are each historical image in the historical image sequence mentioned above.

[0089] Exemplarily, after the model provider obtains the initial image set, for each initial image in the initial image set, the model provider calculates the ratio of the number of power line pixels in the initial image to the total number of pixels in the initial image. If the ratio is greater than or equal to a preset threshold, it means that the grayscale value of the initial image meets the grayscale value condition. If the ratio is less than the preset threshold, it means that the grayscale value of the initial image does not meet the grayscale value condition. Among them, the step of determining the number of power line pixels in the initial image: for any pixel in the initial image, if the difference between the grayscale value of the pixel and the standard grayscale value is less than the grayscale difference threshold, the pixel is determined to be a power line pixel, otherwise, the pixel is not a power line pixel. After traversing all pixels, the number of power line pixels is counted to obtain the number of power line pixels in the initial image.

[0090] In this embodiment, after obtaining the initial image set, the initial image containing the image detail information of the overhead lines of the transmission line is screened out through the gray value condition, so that the image prediction model can be effectively trained based on the image detail information to ensure the effectiveness and accuracy of the model training.

[0091] In a specific embodiment, the interaction process between the target server, the model provider, and each edge node is involved, and the specific steps are as follows:

[0092] Step 1: The target server obtains the data to be detected sent by at least one edge node. Each data to be detected is collected by the collection equipment of the corresponding edge node from a preset position of the overhead line of the transmission line. Each data to be detected includes at least one state data.

[0093] Step 2: The target server determines the data features of each state data, and obtains a reference data feature set corresponding to the overhead line of the transmission line, wherein the reference data feature set includes reference data features corresponding to each data type.

[0094] Step 3: Based on multiple data types involved in the reference data feature set, determine multiple different candidate combinations, each candidate combination defines the data type of each state data; for each candidate combination, calculate the difference between the data feature of the state data and the reference data feature corresponding to the data type based on the data type defined for each state data in the candidate combination, obtain the difference corresponding to the state data, superimpose the differences corresponding to each state data, and obtain the total difference corresponding to the candidate combination; take the candidate combination with the smallest total difference as the target combination, and determine the target data type for each state data based on the target combination.

[0095] Step 4: The target server calls the state prediction model from the model provider. The state prediction model includes a classification network and prediction networks corresponding to each data type. Each state data and the target data type of each state data are input into the state prediction model. Based on the target data type corresponding to each state data, the prediction network corresponding to each state data is determined through the classification network; for each state data, the state data is predicted through the corresponding prediction network to obtain the corresponding prediction result; the operation state of the overhead line of the transmission line is determined through each prediction result.

[0096] The step of determining the state prediction model includes: obtaining each model parameter sent by different edge nodes through the model provider, each model parameter is obtained by the corresponding edge node training the corresponding sub-prediction model according to the training sample of the corresponding data type; for each data type, obtaining at least one model parameter corresponding to the data type through the model provider, fusing at least one model parameter to obtain the corresponding fusion model parameter, and constructing the prediction network corresponding to the data type based on the fusion model parameter; obtaining the state prediction model through the model provider based on each constructed prediction network. In the case where the data type is an image type, the step of determining the training sample of the image type includes: obtaining the initial image set through the edge node corresponding to the image type; screening out the initial image whose gray value meets the gray value condition from the initial image set, and using the screened initial image as the training sample of the image type.

[0097] In this embodiment, by acquiring the data to be detected sent by at least one edge node, each data to be detected is obtained by collecting the preset position of the overhead line of the transmission line by the collection device of the corresponding edge node, and each data to be detected includes at least one state data. That is to say, according to the collection devices of one or more edge nodes, it is possible to complete the comprehensive data collection of the overhead line deployed in the transmission line to ensure the accuracy of the state estimation. Determine the data features of each state data, and obtain the reference data feature set corresponding to the overhead line of the transmission line, the reference data feature set includes the reference data features corresponding to each data type; for each state data, according to the difference between the data features of the state data and each reference data feature, query the target data type matched by the state data. Thus, the data type of each state data is quickly and accurately identified. In this way, based on multiple state data and the target data types of each of the multiple state data, the state prediction model is called, and the operation state of the overhead line of the transmission line can be comprehensively predicted from multiple dimensions with the help of multi-dimensional detailed information, thereby ensuring the accuracy of the state prediction of the overhead line of the transmission line. In addition, through federated learning, the state prediction model is obtained based on the sub-prediction models of the edge nodes. That is, a federated learning strategy is formed between the sub-prediction models and the state prediction model. The model parameters of the sub-prediction models are used to correct and aggregate to form a state prediction model, which is used to accurately predict the operating status of overhead lines of transmission lines. It has good comprehensive performance and engineering practice capabilities, and can ensure the reliability of state perception.

[0098] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0099] Based on the same inventive concept, the embodiment of the present application also provides a device for determining the state of an overhead line of a transmission line for implementing the method for determining the state of an overhead line of a transmission line involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in the embodiments of the device for determining the state of one or more overhead lines of a transmission line provided below can refer to the limitations of the method for determining the state of an overhead line of a transmission line above, and will not be repeated here.

[0100] In an exemplary embodiment, Figure 4 As shown, a device 400 for determining the state of an overhead line of a power transmission line is provided, comprising: a first acquisition module 402, a second acquisition module 404, a query module 406 and a determination module 408, wherein:

[0101] A first acquisition module 402 is used to acquire data to be detected sent by at least one edge node, each data to be detected is obtained by collecting data from a preset position of an overhead line of a transmission line by a collection device corresponding to the edge node, and each data to be detected includes at least one state data;

[0102] A second acquisition module 404 is used to determine data features of each state data, and acquire a reference data feature set corresponding to the overhead line of the transmission line, wherein the reference data feature set includes reference data features corresponding to each data type;

[0103] A query module 406 is used to query, for each state data, the target data type that matches the state data according to the difference between the data feature of the state data and the feature of each reference data;

[0104] The determination module 408 is used to call the state prediction model to determine the operating state of the overhead line of the transmission line based on the multiple state data and the target data types of the multiple state data.

[0105] In one embodiment, the query module 406 is used to calculate the difference between the data feature of each state data and each reference data feature, and query the target data type matched by the state data with the goal of minimizing the difference.

[0106] In one embodiment, the query module 406 is used to determine multiple groups of different candidate combinations based on multiple data types involved in the reference data feature set, and the data type of each state data is defined in each candidate combination; for each group of candidate combinations, according to the data type defined for each state data in the candidate combination, the difference between the data feature of the state data and the reference data feature corresponding to the data type is calculated to obtain the difference corresponding to the state data, and the differences corresponding to each state data are superimposed to obtain the total difference corresponding to the candidate combination; the candidate combination with the smallest total difference is used as the target combination, and based on the target combination, the target data type of each state data is determined respectively.

[0107] In one embodiment, the state prediction model includes a classification network and prediction networks corresponding to each data type, and a determination module 408 is used to input each state data and the target data type of each state data into the state prediction model, and based on the target data type corresponding to each state data, determine the prediction network corresponding to each state data through the classification network; for each state data, perform state prediction on the state data through the corresponding prediction network to obtain a corresponding prediction result; and determine the operating state of the overhead line of the transmission line through each prediction result.

[0108] In one embodiment, the device also includes a construction module, which is used to obtain various model parameters sent by different edge nodes respectively through a model provider, and each model parameter is obtained by the corresponding edge node performing model training on the corresponding sub-prediction model according to the training samples of the corresponding data type; for each data type, at least one model parameter corresponding to the data type is obtained through the model provider, at least one model parameter is fused to obtain the corresponding fused model parameter, and a prediction network corresponding to the data type is constructed based on the fused model parameter; a state prediction model is obtained through the model provider based on each constructed prediction network.

[0109] In one embodiment, when the data type is an image type, the device also includes a screening module for obtaining an initial image set through an edge node corresponding to the image type; screening out initial images whose grayscale values ​​meet the grayscale value conditions from the initial image set, and using the screened initial images as training samples of the image type.

[0110] Each module in the above-mentioned device for determining the state of overhead wires of power transmission lines can be implemented in whole or in part by software, hardware and a combination thereof. Each of the above-mentioned 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 above modules.

[0111] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 5 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. 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 input / output interface of the computer device is used to exchange information between the processor and an external device. The communication 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 state of an overhead line of a transmission line is implemented.

[0112] Those skilled in the art will understand that Figure 5 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.

[0113] In one embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above method embodiments when executing the computer program.

[0114] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0115] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0116] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0117] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method 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 the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. 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). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0118] 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 application.

[0119] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be construed as limiting the scope of the present application. It should be noted that, for a person of ordinary skill in the art, several modifications 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 present application shall be subject to the attached claims.

Claims

1. A method for determining the state of an overhead line of a power transmission line, characterized in that: The method comprises: Acquire data to be detected sent by at least one edge node, each data to be detected is obtained by collecting data from a preset position of an overhead line of a power transmission line by a collection device corresponding to the edge node, and each data to be detected includes at least one state data; Determine data features of each state data, and obtain a reference data feature set corresponding to the overhead line of the transmission line, wherein the reference data feature set includes reference data features corresponding to each data type; For each state data, according to the difference between the data feature of the state data and each reference data feature, query the target data type matched by the state data; Based on a plurality of state data and respective target data types of the plurality of state data, a state prediction model is called to determine the operating state of the overhead line of the transmission line.

2. The method according to claim 1, characterized in that For each state data, according to the difference between the data feature of the state data and each reference data feature, querying the target data type matched by the state data includes: For each state data, the difference between the data feature of the state data and each reference data feature is calculated, and the target data type matched by the state data is queried with the goal of minimizing the difference.

3. The method according to claim 2, characterized in that For each state data, calculating the difference between the data feature of the state data and each reference data feature, and querying the target data type matched by the state data with the goal of minimizing the difference, includes: Based on the multiple data types involved in the reference data feature set, determining multiple groups of different candidate combinations, each group of candidate combinations defining the data type of each state data; For each group of candidate combinations, according to the data type defined for each state data in the candidate combination, the difference between the data feature of the state data and the reference data feature corresponding to the data type is calculated to obtain the difference corresponding to the state data, and the differences corresponding to the various state data are superimposed to obtain the total difference corresponding to the candidate combination; The candidate combination with the smallest total difference is taken as the target combination, and based on the target combination, the target data type of each state data is determined respectively.

4. The method according to claim 1, characterized in that: The state prediction model includes a classification network and prediction networks corresponding to each data type. Based on a plurality of state data and respective target data types of the plurality of state data, calling the state prediction model to determine the operating state of the overhead line of the transmission line includes: Inputting each state data and the target data type of each state data into the state prediction model, and determining the prediction network corresponding to each state data through a classification network based on the target data type corresponding to each state data; For each state data, a state prediction is performed on the state data through a corresponding prediction network to obtain a corresponding prediction result; The operating status of the overhead line of the transmission line is determined through various prediction results.

5. The method according to claim 1, characterized in that The step of determining the state prediction model comprises: Obtain the model parameters sent by different edge nodes through the model provider. Each model parameter is obtained by training the corresponding sub-prediction model through the corresponding edge node according to the training samples of the corresponding data type. For each data type, obtaining at least one model parameter corresponding to the data type through a model provider, fusing at least one model parameter to obtain a corresponding fused model parameter, and constructing a prediction network corresponding to the data type based on the fused model parameter; The state prediction model is obtained by the model provider based on the prediction networks built.

6. The method according to claim 5, characterized in that In the case where the data type is an image type, the step of determining a training sample of the image type includes: Acquire an initial image set through edge nodes corresponding to the image type; Initial images whose grayscale values ​​meet the grayscale value condition are screened out from the initial image set, and the screened out initial images are used as training samples of the image type.

7. A device for determining the state of an overhead line of a power transmission line, characterized in that: The device comprises: A first acquisition module is used to acquire data to be detected sent by at least one edge node, each data to be detected is obtained by collecting data from a preset position of an overhead line of a power transmission line by a collection device corresponding to the edge node, and each data to be detected includes at least one state data; A second acquisition module is used to determine the data features of each state data, and acquire a reference data feature set corresponding to the overhead line of the transmission line, wherein the reference data feature set includes reference data features corresponding to each data type; A query module, configured to query, for each state data, the target data type that matches the state data according to the difference between the data characteristics of the state data and the characteristics of each reference data; The determination module is used to call the state prediction model based on multiple state data and the target data types of the multiple state data to determine the operating state of the overhead line of the transmission line.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. 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 6 are implemented.

10. A computer program product, comprising a computer program, 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 6 are implemented.

Images