Power transmission line icing thickness prediction method and device, computer equipment, readable storage medium and program product

Abstract

The invention relates to a power transmission line icing thickness prediction method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: acquiring target geographic position information of a target prediction moment and a target prediction point; determining current meteorological data and current icing thickness corresponding to the target prediction point; and performing space-time attention processing on the target geographic position information, the current meteorological data and the current icing thickness through a pre-trained power transmission line icing thickness prediction model to obtain a predicted icing thickness corresponding to the target prediction point at the target prediction moment. By adopting the method, the icing thickness prediction accuracy can be improved.

Description

Technical Field

[0001] This application relates to the technical field of smart grids, and in particular, to a method, device, computer device, computer-readable storage medium, and computer program product for predicting the icing thickness of transmission lines. Background Art

[0002] Icing on transmission lines is one of the main natural disasters threatening the safe operation of the power grid. Its formation requires the combined action of low temperature (usually below 0°C), high humidity, and appropriate wind speed. When supercooled water droplets in the air come into contact with the surface of the wire, they will quickly freeze and continuously accumulate to form an ice layer. When the ice layer thickness exceeds the design threshold, the wire tension rises sharply, which may cause the fittings to break, the tower to overturn, or even the line to collapse. In related technologies, icing observation mainly relies on manual inspections. Developing high-precision icing growth prediction technologies to achieve proactive prevention before disasters is of great practical significance for ensuring the resilience of the power grid.

[0003] In related technologies, icing prediction technologies mainly rely on data from a single monitoring station, ignoring the geographical distribution and regional meteorological synergy effects, resulting in poor prediction accuracy. Summary of the Invention

[0004] Based on this, in order to solve the above technical problems, it is necessary to provide a method, device, computer device, computer-readable storage medium, and computer program product for predicting the icing thickness of transmission lines that can improve prediction accuracy.

[0005] In a first aspect, this application provides a method for predicting the icing thickness of transmission lines. The method includes:

[0006] Obtain the target geographical location information of the target prediction time and the target prediction point;

[0007] Determine the current meteorological data and the current icing thickness corresponding to the target prediction point;

[0008] Through a pre-trained transmission line icing thickness prediction model, perform spatio-temporal attention processing on the target geographical location information, the current meteorological data, and the current icing thickness to obtain the predicted icing thickness corresponding to the target prediction point at the target prediction time.

[0009] In one embodiment, before obtaining the predicted ice accretion thickness corresponding to the target prediction point at the target prediction time through a pre-trained transmission line ice accretion thickness prediction model by performing spatio-temporal attention processing on the target geographical location information, current meteorological data, and current ice accretion thickness, the method further includes: obtaining the monitored geographical location information and monitored meteorological feature data of each monitoring point, and generating a joint encoding matrix according to the monitored geographical location information and monitored meteorological feature data; the joint encoding matrix is used to represent the correlation between the monitored geographical location, meteorological feature data, and ice accretion thickness; determining the geographical topological correlation matrix between each monitoring point and its adjacent monitoring points according to the monitored geographical location, and generating a meteorological feature similarity matrix according to the geographical topological correlation matrix; generating a meteorological time series matrix according to the geographical topological correlation matrix and the meteorological feature similarity matrix; generating a model training matrix according to the joint encoding matrix, the meteorological feature similarity matrix, and the meteorological time series matrix, and training to generate a transmission line ice accretion thickness prediction model according to the model training matrix.

[0010] In one of the embodiments, the monitored geographical location information includes monitored longitude and latitude data and monitored altitude data; generating a joint encoding matrix according to the monitored geographical location information and meteorological feature data includes: generating a geographical feature vector according to the monitored longitude and latitude data and monitored altitude data; generating a meteorological feature vector according to the monitored meteorological feature data; using the ice accretion thickness of each monitoring point as a query vector to perform cross-modal attention calculation on the geographical feature vector and the meteorological feature vector to determine the correlation between the monitored geographical location, monitored meteorological feature data, and ice accretion thickness, and generating a joint encoding matrix according to the correlation.

[0011] In an optional embodiment, generating a geographical feature vector according to the monitored longitude and latitude data and monitored altitude data includes: generating a sine encoding matrix corresponding to the monitored longitude and latitude data according to the monitored longitude and latitude data; generating a terrain feature vector corresponding to the monitored altitude data according to the monitored altitude data; generating a geographical feature vector according to the sine encoding matrix and the terrain feature vector.

[0012] In one embodiment, determining the geographical topological correlation matrix between each monitoring point and its adjacent monitoring points according to the monitored geographical location, and generating a meteorological feature similarity matrix according to the geographical topological correlation matrix includes: determining the geographical topological correlation degree between each monitoring point according to the monitored geographical location information of each monitoring point, and generating a geographical topological correlation matrix according to the geographical topological correlation degree; using the geographical topological correlation matrix as a prior condition and the ice accretion thickness as a query vector to determine the similarity between the meteorological feature data of the monitoring point and the meteorological feature data of its adjacent monitoring points, and generating a meteorological feature similarity matrix.

[0013] In an alternative embodiment, a meteorological time series matrix is generated according to the geographical topology association matrix and the meteorological feature similarity matrix, including: obtaining adjacent monitoring points corresponding to each monitoring point from the geographical topology association matrix; determining the average ice coating thickness according to the ice coating thickness of each monitoring point and the ice coating thickness of the corresponding adjacent monitoring points; obtaining meteorological time series features from the meteorological feature similarity matrix, and re-calibrating the weights of the meteorological time series features according to the meteorological time series features and the average ice coating thickness; generating a meteorological time series matrix according to the meteorological time series features after weight re-calibration.

[0014] In a second aspect, the present application also provides a device for predicting the ice coating thickness of a transmission line, including:

[0015] A prediction information acquisition module, configured to acquire the target geographical location information of the target prediction time and the target prediction point;

[0016] A prediction data acquisition module, configured to determine the current meteorological data and the current ice coating thickness corresponding to the target prediction point;

[0017] A prediction result determination module, configured to perform spatio-temporal attention processing on the target geographical location information, the current meteorological data, and the current ice coating thickness through a pre-trained ice coating thickness prediction model for a transmission line, so as to obtain the predicted ice coating thickness corresponding to the target prediction point at the target prediction time.

[0018] In a third aspect, the present application also provides a computer device, including a memory and a processor, where the memory stores a computer program, and when the processor executes the computer program, the steps of the above method embodiments are implemented.

[0019] In a fourth aspect, the present application also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above method embodiments are implemented.

[0020] In a fifth aspect, the present application also provides a computer program product, including a computer program, and when the computer program is executed by a processor, the steps of the above method embodiments are implemented.

[0021] The above method, device, computer device, computer-readable storage medium, and computer program product for predicting the ice coating thickness of a transmission line utilize a pre-trained ice coating thickness prediction model for a transmission line to perform spatio-temporal attention processing on the target geographical location information, the current meteorological data, and the current ice coating thickness, combine the geographical location information with the meteorological data, enhance the fusion of static features (geographical information) and dynamic features (meteorological data), and effectively improve the accuracy of ice coating thickness prediction. Description of the Drawings

[0022] To more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following will briefly introduce the accompanying drawings required for the description of the embodiments of the present application or related technologies. Obviously, the accompanying drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings.

[0023] Figure 1 It is an application environment diagram of the transmission line ice thickness prediction method in an embodiment;

[0024] Figure 2 It is a schematic flowchart of the transmission line ice thickness prediction method in an embodiment;

[0025] Figure 3 It is a schematic flowchart of the transmission line ice thickness prediction model generation method in an embodiment;

[0026] Figure 4 It is a structural block diagram of the transmission line ice thickness prediction device in an embodiment;

[0027] Figure 5 It is an internal structure diagram of a computer device in an embodiment. Detailed implementation manners

[0028] In order to make the purpose, technical solutions and advantages of the present application clearer, the following further details the present application 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.

[0029] The transmission line ice thickness prediction method provided by the embodiments of the present application can be applied to, for example Figure 1In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or can be placed on the cloud or other network servers. The server 104 obtains the target prediction time and the target geographical location information of the target prediction point transmitted by the terminal 102, and determines the current meteorological data and the current ice coating thickness corresponding to the target prediction point; through a pre-trained transmission line ice coating thickness prediction model, performs spatio-temporal attention processing on the target geographical location information, the current meteorological data, and the current ice coating thickness, and obtains the predicted ice coating thickness corresponding to the target prediction point at the target prediction time. Among them, the terminal 102 can be, but is not limited to, various personal computers, laptop computers, smart phones, tablet computers, Internet of Things devices, and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, projection devices, etc. The portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The head-mounted device can be a virtual reality (VR) device, an augmented reality (AR) device, smart glasses, etc. The server 104 can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services.

[0030] In an exemplary embodiment, as Figure 2 shown, a method for predicting the ice coating thickness of a transmission line is provided. Taking the server in Figure 1 as an example, the method includes the following steps 202 to step 206. Among them:

[0031] Step 202, obtain the target prediction time and the target geographical location information of the target prediction point.

[0032] Among them, the target prediction time is used to represent the time information for which the ice coating thickness prediction is expected. Optionally, the target prediction time can be a certain time point or a time period.

[0033] Among them, the target geographical location information is used to represent the location information for which the ice coating thickness prediction is expected. Optionally, the target geographical location information includes the target longitude and latitude and the target altitude.

[0034] In one embodiment, the target prediction time and the target geographical location information of the target prediction point can be input by the user using the terminal and transmitted by the terminal to the server, so that after the server receives the target prediction time and the target geographical location information of the target prediction point, it can predict the ice coating thickness of the transmission line.

[0035] In one embodiment, information selection software can be set on the terminal, and check boxes for determining the target prediction time and the target geographical location information of the target prediction point can be set on the information selection software, for selecting the desired prediction time and the desired prediction point in the check boxes. The information selection software determines the target prediction time according to the desired prediction time selected by the user, and determines the target geographical location information of the target prediction point according to the desired prediction point selected by the user.

[0036] Optionally, a text box can also be set on the terminal, and the user directly inputs the target prediction time and the target geographical location information of the target prediction point in the text box. So that after the terminal receives the target prediction time and the target geographical location information of the target prediction point input by the user, it sends the received target prediction time and the target geographical location information of the target prediction point to the server, and the server predicts the ice coating thickness based on the received target prediction time and the target geographical location information of the target prediction point.

[0037] Step 204, determine the current meteorological data and the current ice coating thickness corresponding to the target prediction point.

[0038] Among them, the meteorological data is used to represent the data reflecting the weather conditions at the current prediction point. Optionally, the meteorological data may include but is not limited to temperature, humidity, wind speed, wind direction, ice coating thickness, and liquid water content, etc.

[0039] In one embodiment, after determining the target prediction point, the current meteorological data of the target prediction point and the current ice coating thickness of the transmission line corresponding to the target prediction point can be obtained.

[0040] In one of the embodiments, the current meteorological data can be obtained from the official or professional website corresponding to the target prediction point.

[0041] In an alternative embodiment, for obtaining the current ice coating thickness of the target prediction point, the monitoring device installed on the transmission line can be used to obtain the ice coating thickness data in real time, or the staff can directly measure the ice coating thickness through on-site observation and measuring tools.

[0042] Step 206, perform spatio-temporal attention processing on the target geographical location information, the current meteorological data, and the current ice coating thickness through a pre-trained transmission line ice coating thickness prediction model, to obtain the predicted ice coating thickness corresponding to the target prediction point at the target prediction time.

[0043] In one embodiment, the transmission line ice coating thickness prediction model can be pre-trained, so that when prediction is required, it predicts based on the target geographical location information, the current meteorological data, and the current ice coating thickness, to obtain the predicted ice coating thickness corresponding to the target prediction point at the target prediction time.

[0044] Among them, spatio-temporal attention processing is used to synthesize various meteorological parameters included in meteorological data (such as temperature, humidity, wind speed, wind direction, ice coating thickness, and liquid water content) to achieve the prediction of ice coating thickness.

[0045] In the above method for predicting the ice coating thickness of a transmission line, a pre-trained prediction model for the ice coating thickness of a transmission line is used to perform spatio-temporal attention processing on the target geographical location information, current meteorological data, and current ice coating thickness, so as to combine the geographical location information with the meteorological data, improve the fusion of static features (geographical information) and dynamic features (meteorological data), and effectively improve the accuracy of ice coating thickness prediction.

[0046] In an exemplary embodiment, before using the prediction model for the ice coating thickness of a transmission line to make a prediction, the prediction model for the ice coating thickness of a transmission line can be trained first, including:

[0047] Step 302, obtain the monitored geographical location information and monitored meteorological feature data of each monitoring point, and generate a joint coding matrix according to the monitored geographical location information and monitored meteorological feature data.

[0048] Among them, the joint coding matrix is used to represent the correlation relationship between the monitored geographical location, meteorological feature data, and ice coating thickness.

[0049] Among them, the monitoring points are pre-set and are used to observe the ice coating thickness of the transmission line under different monitored geographical locations and / or different monitored meteorological feature data.

[0050] In one embodiment, there can be multiple monitoring points. Optionally, different monitoring points can have different monitored geographical locations and different monitored meteorological feature data.

[0051] In one of the embodiments, the monitored geographical location information includes but is not limited to monitored longitude and latitude data, monitored altitude data, etc.

[0052] In one of the embodiments, the monitored meteorological feature data can include but is not limited to monitored temperature, monitored humidity, monitored wind speed, monitored wind direction, monitored ice coating thickness, and monitored liquid water content, etc.

[0053] In one embodiment, after obtaining the monitored meteorological feature data, the monitored meteorological feature data can be preprocessed so that the monitored meteorological feature data meets the data processing requirements.

[0054] Exemplarily, the preprocessing of the monitored meteorological feature data can be implemented by adopting a sliding window normalization processing method. Optionally, the data segments can be intercepted by sliding along the time axis with a step of 1 hour and a time window length of 24 hours.

[0055] In an exemplary embodiment, each sliding window may include at least six categories of monitored meteorological feature data.

[0056] In one embodiment, the monitored meteorological feature data within each window may be standardized. Exemplarily, the Z-score normalization process may be adopted for the monitored meteorological feature data within each window to obtain the processed and standardized meteorological feature data.

[0057] Optionally, the Z-score normalization formula is as shown in formula (1):

[0058] (1)

[0059] Wherein, is the feature mean within the window, is the standard deviation.

[0060] In one of the embodiments, after normalizing the monitored meteorological feature data within each window to obtain the standardized meteorological feature data, a meteorological standardized feature matrix may be generated based on the standardized meteorological feature data, so as to convert the monitored meteorological feature data from the initial dimension [N×M×6] (where N = the number of monitoring points, M = the total historical duration, and 6 is the number of features) into a standardized feature matrix [N×T×D], where T = 24 is the time window length and D = 6 is the feature dimension.

[0061] In an optional embodiment, when the monitored meteorological feature data includes liquid water content, different processing methods may be adopted for different liquid water contents. Optionally, when freezing rain weather is detected, a logarithmic transformation is performed on the liquid water content to amplify its numerical difference.

[0062] In an optional embodiment, when the temperature within the window > 5°, the window may be directly marked as a non-icing risk period to reduce invalid calculations.

[0063] In one of the embodiments, a joint coding matrix is generated based on the monitored geographical location information and meteorological feature data, including: generating a geographical feature vector according to the monitored longitude and latitude data and monitored altitude data; generating a meteorological feature vector according to the monitored meteorological feature data; using the ice thickness of each monitoring point as a query vector to perform cross-modal attention calculation on the geographical feature vector and the meteorological feature vector to determine the association relationship between the monitored geographical location, the monitored meteorological feature data, and the ice thickness, and generating a joint coding matrix according to the association relationship.

[0064] Optionally, a geographical feature vector can be generated based on the monitored longitude and latitude data and the monitored altitude data, which may include: generating a sine encoding matrix corresponding to the monitored longitude and latitude data according to the monitored longitude and latitude data; generating a terrain feature vector corresponding to the monitored altitude data according to the monitored altitude data; generating a geographical feature vector based on the sine encoding matrix and the terrain feature vector.

[0065] In one embodiment, a sine position encoding can be obtained by processing the monitored longitude and latitude data using a geographical encoder, and the specific calculation formula is shown in Formula (2):

[0066] (2)

[0067] where d = 64, which is the encoding dimension.

[0068] Optionally, a sine encoding matrix can be generated based on the calculated sine position encoding. Exemplarily, a 64-dimensional sine encoding matrix is generated.

[0069] In one embodiment, the altitude data can be processed using a Gaussian kernel function to obtain the altitude elevation, and a terrain feature vector can be obtained based on the obtained altitude elevation. For example, an 8-dimensional vector is generated.

[0070] Optionally, the bandwidth of the Gaussian kernel function can be dynamically adjusted according to the regional terrain undulation, as shown in Formula (3):

[0071] (3)

[0072] where h is the altitude of the current monitoring point, are 8 preset altitude reference points, is the bandwidth parameter.

[0073] In one embodiment, after generating the sine encoding feature vector and the terrain feature vector, the sine encoding feature vector and the terrain feature vector are fused to obtain a geographical feature vector. Exemplarily, the generated geographical feature vector can be [N, 8], where N is the number of monitoring points and 8 is the dimension of the terrain feature vector of each monitoring point.

[0074] In an optional embodiment, the geographical feature vector can be further processed to obtain a spatial position embedding vector. Optionally, a geographical GRU (gated recurrent unit) can be used to further process the geographical feature vector to obtain a spatial position embedding vector. Exemplarily, the geographical GRU is a unidirectional structure.

[0075] In one embodiment, a meteorological GRU can be used to process the preprocessed meteorological standardized feature matrix, extract the meteorological temporal features in the meteorological standardized feature matrix, and generate a meteorological feature vector based on the extracted meteorological temporal features. Optionally, the meteorological GRU can be a bidirectional GRU.

[0076] In one embodiment, the output of the meteorological GRU can be as shown in formula (4):

[0077] (4)

[0078] Wherein, is the meteorological feature input within the time window, with a dimension of (T = 24, D = 6); is the weight parameter of the GRU. The weight dimension of the bidirectional GRU is 128, and the weight dimension of the unidirectional GRU is 64.

[0079] Optionally, the association relationship can be characterized by a correlation weight.

[0080] Among them, cross-modal attention calculation is a technology for processing multi-modal data (such as text, images, videos, geographical information, meteorological data, etc.), aiming to capture the correlation and interaction between different modal data, and dynamically adjust the contribution weights of different modal data to the target task through the attention mechanism.

[0081] Optionally, the cross-modal attention mechanism may include the following key parts:

[0082] Query: Represents the main focus of the current task, usually the feature representation of the target modality.

[0083] Key and Value: Represent the feature representations of other modalities. The key is used to calculate the similarity with the query, and the value is used to generate the final weighted feature representation.

[0084] Attention Scores: Obtained by calculating the similarity (such as dot product, cosine similarity, etc.) between the query and the key, representing the contribution weights of different modal data to the target task.

[0085] Exemplarily, the ice coating thickness can be used as the query vector, and the geographical features and meteorological features can be used as the key and value to determine the correlation weight of the geographical features to the ice coating thickness and the correlation weight of the meteorological features to the ice coating thickness, and generate a joint coding matrix according to the correlation weights.

[0086] Step 304: Determine the geographical topological association matrix between each monitoring point and its adjacent monitoring points according to the monitored geographical location, and generate a meteorological feature similarity matrix according to the geographical topological association matrix.

[0087] Optionally, the geographical topological correlation degree between monitoring points can be determined according to the geographical location information of each monitoring point, and a geographical topological correlation matrix can be generated based on the geographical topological correlation degree; taking the geographical topological correlation matrix as a prior condition and the ice coating thickness as a query vector, the similarity between the meteorological characteristic data of a monitoring point and the meteorological characteristic data of adjacent monitoring points can be determined, and a meteorological characteristic similarity matrix can be generated.

[0088] In one embodiment, the calculation formula of the geographical topological correlation degree between monitoring points can be as shown in formula (5):

[0089] (5)

[0090] Wherein, is the elevation difference, is the smoothing coefficient, is the sine coding matrix of monitoring point i, is the sine coding matrix of monitoring point j.

[0091] Optionally, after determining the geographical topological correlation degree between each monitoring point, a spatial correlation matrix A can be generated according to the calculated geographical topological correlation degree between each monitoring point.

[0092] In an exemplary embodiment, the spatial correlation matrix A can be used as a prior condition, combined with the attention score ( ), taking the ice coating thickness of the current monitoring point as a query vector, the similarity between the meteorological characteristic data of the current monitoring point and the meteorological characteristic data of adjacent monitoring points can be determined, and a meteorological similarity matrix can be generated according to the obtained similarity between the meteorological characteristic data of the current monitoring point and the meteorological characteristic data of adjacent monitoring points.

[0093] In one embodiment, a multi-head attention layer can be stacked on the hidden state layer of the meteorological GRU, and the attention layer can be used to determine the similarity between the meteorological characteristic data of the current monitoring point and the meteorological characteristic data of adjacent monitoring points.

[0094] Optionally, the formula of the multi-head attention layer can be as shown in formula (6):

[0095] (6)

[0096] Wherein, , , are the query, key, and value matrices after linear transformation, is the spatial correlation matrix. The number of attention heads is set to 8, and the dimension of each head , and the output feature dimension remains unchanged at 256.

[0097] Step 306: Generate a meteorological time series matrix based on the geographical topology association matrix and the meteorological feature similarity matrix.

[0098] Optionally, the adjacent monitoring points corresponding to each monitoring point can be obtained from the geographical topology association matrix; the average ice coating thickness can be determined according to the ice coating thickness of each monitoring point and the ice coating thickness of the corresponding adjacent monitoring points; the meteorological time series features can be obtained from the meteorological feature similarity matrix, and the meteorological time series features can be reweighted according to the meteorological time series features and the average ice coating thickness; a meteorological time series matrix can be generated according to the reweighted meteorological time series features.

[0099] In one embodiment, the geographical topology association degree in the spatial association matrix A is used to determine the adjacent monitoring points of each monitoring point.

[0100] In one of the embodiments, the ice coating thickness of each monitoring point and the adjacent monitoring points is obtained, and the average ice coating thickness is determined according to the ice coating thickness.

[0101] In one embodiment, the time window for reweighting can be determined first, and the average ice coating thickness of each monitoring point and the adjacent monitoring points within the time window can be determined. Exemplarily, if the time window is [t - 6, t], then the average ice coating thickness of each monitoring point and the adjacent monitoring points within the time window [t - 6, t] can be .

[0102] In one of the embodiments, a bidirectional GRU can be used to obtain the meteorological time series features from the meteorological feature similarity matrix, and the meteorological time series features can be reweighted based on the obtained meteorological time series features and the average ice coating thickness to determine the importance of the factors affecting the ice coating thickness. For example, before reweighting, temperature may be an important factor affecting the ice coating thickness, and after reweighting, it may be humidity.

[0103] Optionally, the formula for reweighting the meteorological time series features can be as shown in formula (7):

[0104] (7)

[0105] Where, , is a learnable parameter, is the average ice coating thickness of the neighborhood monitoring points.

[0106] Optionally, the meteorological time series features are reweighted to obtain the reweighted meteorological time series features, and a meteorological time series matrix is generated according to the reweighted meteorological time series features.

[0107] Step 308: Generate a model training matrix based on the joint coding matrix, the meteorological feature similarity matrix, and the meteorological time series matrix, and train a transmission line icing thickness prediction model based on the model training matrix.

[0108] In one embodiment, the formula used for training the model can be as shown in Formula (8):

[0109] (8)

[0110] Wherein, is the weight mixing coefficient reflecting different growth modes (such as rapid growth, slow growth); is the growth amplitude scale parameter reflecting the icing thickness. Optionally, ; is the shape parameter controlling the growth rate and right-skewness degree of the icing thickness. Optionally, .

[0111] In one of the embodiments, when training the generation model, the confidence level of the output result can also be trained. For example, data with a confidence level of 95% can be output, and the mean value of the icing thickness in the prediction time period can also be output. Optionally, the formula used can be as shown in Formula (9):

[0112] (9)

[0113] Wherein, .

[0114] In one embodiment, when training, the loss function can adopt the joint optimization of the CRPS continuous ranked probability score and the KL divergence, and the formula is as shown in (10):

[0115] (10)

[0116] Wherein, , , is the kernel density estimation distribution of the observed data. A meteorological mutation simulator is introduced, and temperature steps of ±5°C and wind speed pulses of ±10 m / s are randomly injected. The probability of the perturbation occurrence is set to 15%.

[0117] In the above embodiments, a prediction model for the icing thickness of a transmission line is pre-generated, so that the prediction model for the icing thickness of the transmission line can be called when needed to predict the predicted icing thickness corresponding to the target prediction point at the target prediction time. Moreover, in the above embodiments, during the training process of the transmission line icing thickness prediction model, the geographical location information and meteorological characteristic data of multiple monitoring points are used to improve the diversity of the training data and the accuracy of the trained model, thereby improving the accuracy of the predicted icing thickness result obtained based on the model. Further, in the above embodiments, the geographical topological correlation degree between the monitoring points is also considered, the monitoring points and the adjacent monitoring points are associated by using the geographical topological structure, and the correlation between the meteorological characteristics of the monitoring points and the adjacent monitoring points is considered, so as to combine the geographical characteristics, meteorological characteristics and the similarity information between the monitoring points, effectively improving the accuracy of model training and thus the accuracy of icing thickness prediction.

[0118] In one embodiment, when using the transmission line icing thickness prediction model to perform spatio-temporal attention processing on the target geographical location information, the current meteorological data, and the current icing thickness to obtain the predicted icing thickness corresponding to the target prediction point at the target prediction time, it is also possible to process the target geographical location information, the current meteorological data, the current icing thickness, etc., and then use the transmission line icing thickness prediction model to perform spatio-temporal attention processing on the target geographical location information, the current meteorological data, and the current icing thickness to obtain the predicted icing thickness corresponding to the target prediction point at the target prediction time. The specific implementation process is similar to steps 302 to 308 and will not be elaborated here.

[0119] It should be understood that although the steps in the flowcharts involved in the above embodiments are shown in sequence according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless there is a clear description in this article, the execution of these steps has no strict order limit, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be executed alternately or alternately with at least a part of other steps or steps or stages in other steps.

[0120] Based on the same inventive concept, an embodiment of the present application further provides a transmission line ice thickness prediction device for implementing the transmission line ice thickness prediction method involved above. The solution provided by this device to solve the problem is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the transmission line ice thickness prediction device provided below can refer to the limitations on the transmission line ice thickness prediction method in the above text, and will not be repeated here.

[0121] In an exemplary embodiment, as Figure 4 shown, a transmission line ice thickness prediction device 400 is provided, including: a prediction information acquisition module 402, a prediction data acquisition module 404, and a prediction result determination module 406, where:

[0122] The prediction information acquisition module 402 is configured to acquire the target geographical location information of the target prediction time and the target prediction point;

[0123] The prediction data acquisition module 404 is configured to determine the current meteorological data and the current ice thickness corresponding to the target prediction point;

[0124] The prediction result determination module 406 is configured to perform spatio-temporal attention processing on the target geographical location information, the current meteorological data, and the current ice thickness through a pre-trained transmission line ice thickness prediction model, so as to obtain the predicted ice thickness corresponding to the target prediction point at the target prediction time.

[0125] In one embodiment, through a pre-trained transmission line ice thickness prediction model, the device further includes:

[0126] A model generation module, configured to acquire the monitoring geographical location information and the monitoring meteorological feature data of each monitoring point, and generate a joint coding matrix according to the monitoring geographical location information and the monitoring meteorological feature data; the joint coding matrix is used to represent the association relationship between the monitoring geographical location, the meteorological feature data, and the ice thickness; according to the monitoring geographical location, determine the geographical topological association matrix between each monitoring point and its adjacent monitoring points, and generate a meteorological feature similarity matrix according to the geographical topological association matrix; generate a meteorological time series matrix according to the geographical topological association matrix and the meteorological feature similarity matrix; generate a model training matrix according to the joint coding matrix, the meteorological feature similarity matrix, and the meteorological time series matrix, and train and generate a transmission line ice thickness prediction model according to the model training matrix.

[0127] In one embodiment, monitoring the geographical location information includes monitoring longitude and latitude data and monitoring altitude data; the model generation module is further configured to generate a geographical feature vector according to the monitored longitude and latitude data and the monitored altitude data; generate a meteorological feature vector according to the monitored meteorological feature data; use the ice thickness at each monitoring point as a query vector to perform cross-modal attention calculation on the geographical feature vector and the meteorological feature vector to determine the association relationship between the monitored geographical location, the monitored meteorological feature data and the ice thickness, and generate a joint coding matrix according to the association relationship.

[0128] In an alternative embodiment, the model generation module is further configured to generate a sine coding matrix corresponding to the monitored longitude and latitude data according to the monitored longitude and latitude data; generate a terrain feature vector corresponding to the monitored altitude data according to the monitored altitude data; generate a geographical feature vector according to the sine coding matrix and the terrain feature vector.

[0129] In one embodiment, the model generation module is further configured to determine the geographical topological correlation degree between each monitoring point according to the monitored geographical location information of each monitoring point, and generate a geographical topological correlation matrix according to the geographical topological correlation degree; use the geographical topological correlation matrix as a prior condition and the ice thickness as a query vector to determine the similarity between the meteorological feature data of the monitoring point and the meteorological feature data of the adjacent monitoring point, and generate a meteorological feature similarity matrix.

[0130] In one embodiment, the model generation module is further configured to obtain the adjacent monitoring points corresponding to each monitoring point from the geographical topological correlation matrix; determine the average ice thickness according to the ice thickness of each monitoring point and the ice thickness of the corresponding adjacent monitoring point; obtain the meteorological time series feature from the meteorological feature similarity matrix, and perform weight recalibration on the meteorological time series feature according to the meteorological time series feature and the average ice thickness; generate a meteorological time series matrix according to the meteorological time series feature after weight recalibration.

[0131] Each module in the above transmission line ice thickness prediction device can be implemented in whole or in part by software, hardware and their combination. The above modules can be embedded in the processor in the computer device in the form of hardware or be independent of the processor, or can be stored in the memory in the computer device in the form of software, so as to facilitate the processor to call and execute the operations corresponding to the above modules.

[0132] In an exemplary embodiment, a computer device is provided. The computer device can be a server, and its internal structure diagram can be as Figure 5As shown in the figure. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, 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. 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. The input / output interface of the computer device is used to exchange information between the processor and external devices. 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, it implements a method for predicting the icing thickness of a transmission line.

[0133] Those skilled in the art can understand that Figure 5 the structure shown in the figure is only a block diagram of some structures related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0134] In an exemplary embodiment, a computer device is provided, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the steps of the above method embodiments are implemented.

[0135] In an embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by the processor, the steps of the above method embodiments are implemented.

[0136] In an embodiment, a computer program product is provided, including a computer program. When the computer program is executed by the processor, the steps of the above method embodiments are implemented.

[0137] 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 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 need to comply with relevant regulations.

[0138] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. 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 methods. Among them, any reference to a 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), magnetoresistive 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. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logic devices, data processing logics based on quantum computing, artificial intelligence (AI) processors, etc., without limitation.

[0139] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, 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, it should be considered as the scope recorded in the present application.

[0140] The above embodiments merely represent several implementation manners of the present application. The description thereof is relatively specific and detailed, but it should not be construed as a limitation to the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all fall within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the appended claims.

Claims

1. A method for predicting the icing thickness of a transmission line, characterized in that, The method includes: Obtaining the target geographical location information of the target prediction moment and the target prediction point; Determining the current meteorological data and the current ice coating thickness corresponding to the target prediction point; Performing spatio-temporal attention processing on the target geographical location information, the current meteorological data, and the current ice coating thickness through the pre-trained ice coating thickness prediction model for transmission lines, to obtain the predicted ice coating thickness corresponding to the target prediction point at the target prediction moment.

2. The method according to claim 1, characterized in that, Before performing spatio-temporal attention processing on the target geographical location information, the current meteorological data, and the current ice coating thickness through the pre-trained ice coating thickness prediction model for transmission lines to obtain the predicted ice coating thickness corresponding to the target prediction point at the target prediction moment, the method further includes: Obtaining the monitored geographical location information and monitored meteorological feature data of each monitoring point, and generating a joint coding matrix according to the monitored geographical location information and the monitored meteorological feature data; the joint coding matrix is used to characterize the correlation between the monitored geographical location, the meteorological feature data, and the ice coating thickness; Determining the geographical topological correlation matrix between each monitoring point and its adjacent monitoring points according to the monitored geographical location, and generating a meteorological feature similarity matrix according to the geographical topological correlation matrix; Generating a meteorological time series matrix according to the geographical topological correlation matrix and the meteorological feature similarity matrix; Generating a model training matrix according to the joint coding matrix, the meteorological feature similarity matrix, and the meteorological time series matrix, and training to generate the ice coating thickness prediction model for transmission lines according to the model training matrix.

3. The method according to claim 2, wherein The monitored geographical location information includes monitored longitude and latitude data and monitored altitude data; generating the joint coding matrix according to the monitored geographical location information and the meteorological feature data includes: Generating a geographical feature vector according to the monitored longitude and latitude data and the monitored altitude data; Generating a meteorological feature vector according to the monitored meteorological feature data; Using the ice coating thickness of each monitoring point as a query vector, performing cross-modal attention calculation on the geographical feature vector and the meteorological feature vector to determine the correlation between the monitored geographical location, the monitored meteorological feature data, and the ice coating thickness, and generating the joint coding matrix according to the correlation.

4. The method according to claim 3, wherein Generating the geographical feature vector according to the monitored longitude and latitude data and the monitored altitude data includes: Generating a sine coding matrix corresponding to the monitored longitude and latitude data according to the monitored longitude and latitude data; Generating a terrain feature vector corresponding to the monitored altitude data according to the monitored altitude data; Generating the geographical feature vector according to the sine coding matrix and the terrain feature vector.

5. The method according to claim 2, wherein Determining the geographical topological correlation matrix between each monitoring point and its adjacent monitoring points according to the monitored geographical location, and generating a meteorological feature similarity matrix according to the geographical topological correlation matrix includes: Determining the geographical topological correlation degree between each monitoring point according to the monitored geographical location information of each monitoring point, and generating a geographical topological correlation matrix according to the geographical topological correlation degree; Taking the geographical topology association matrix as a prior condition and the ice coating thickness as a query vector, determine the similarity between the meteorological characteristic data of the monitoring points and the meteorological characteristic data of adjacent monitoring points, and generate the meteorological characteristic similarity matrix.

6. The method according to claim 5, characterized in that, Generating a meteorological time series matrix according to the geographical topology association matrix and the meteorological characteristic similarity matrix, including: Obtain the adjacent monitoring points corresponding to each of the monitoring points from the geographical topology association matrix; Determine the average ice coating thickness according to the ice coating thickness of each of the monitoring points and the ice coating thickness of the corresponding adjacent monitoring points; Obtain meteorological time series characteristics from the meteorological characteristic similarity matrix, and re-calibrate the weights of the meteorological time series characteristics according to the meteorological time series characteristics and the average ice coating thickness; Generate a meteorological time series matrix according to the meteorological time series characteristics after weight re-calibration.

7. An ice thickness prediction device for a transmission line, characterized in that, The device includes: A prediction information acquisition module, configured to acquire the target geographical location information of the target prediction time and the target prediction point; A prediction data acquisition module, configured to determine the current meteorological data and the current ice coating thickness corresponding to the target prediction point; A prediction result determination module, configured to perform spatio-temporal attention processing on the target geographical location information, the current meteorological data, and the current ice coating thickness through the pre-trained transmission line ice coating thickness prediction model, to obtain the predicted ice coating thickness corresponding to the target prediction point at the target prediction time.

8. A computer device, comprising a memory and a processor, the memory storing 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 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 the 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 the processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images