Global safety early warning method and system based on local power transmission line wind vibration frequency monitoring

Through the global safety early warning method based on local transmission line wind vibration frequency monitoring, the problems of high deployment costs and poor information utilization in large-scale transmission lines are solved, and the timeliness and accuracy of global safety early warning is achieved, reducing monitoring costs.

CN118917487APending Publication Date: 2024-11-08GUANGDONG POWER GRID CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411112940.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-14
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

The existing transmission line wind vibration frequency monitoring system is costly and difficult to deploy in large-scale transmission lines, and it cannot achieve effective use of information and early warning, especially when the terrain of different transmission towers is different, it is difficult to apply to Large-scale transmission scenarios.

Method used

Using the global safety warning method based on local transmission line wind vibration frequency monitoring, the real-time monitoring data of multiple preset positions of the transmission line is obtained, and input it into the wind vibration frequency prediction model to generate the wind vibration of each position in the next preset time period. Frequency change prediction data set, and early warning level is calculated based on the prediction data and importance coefficient to achieve global security warning.

Benefits of technology

It reduces the cost of sensor deployment and wind vibration monitoring, improves the timeliness of early warning, and can deploy sensors at designated locations to achieve global monitoring and early warning of transmission lines, and predicts global information through local information, improving the accuracy of prediction.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118917487A_ABST
    Figure CN118917487A_ABST
Patent Text Reader

Abstract

The invention provides a global safety early warning method and system based on local power transmission line wind vibration frequency monitoring. The method comprises the following steps: acquiring real-time monitoring data corresponding to a plurality of preset positions of a power transmission line; inputting the plurality of pieces of real-time monitoring data into a preset wind vibration frequency prediction model, so that the wind vibration frequency prediction model generates a prediction data set of wind vibration frequency change of each position in the power transmission line in a next preset time period according to the plurality of pieces of real-time monitoring data and a preset power transmission line topological graph; performing evaluation according to the prediction data set and a preset threshold value to obtain a safety level of each position in the power transmission line; performing comprehensive calculation according to the safety level of each position in the power transmission line and a corresponding preset importance coefficient to obtain an early warning level of each position in the power transmission line; and performing safety early warning according to the early warning level, thereby realizing global early warning of the power transmission line, and improving early warning timeliness while reducing deployment cost.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of wind vibration monitoring and early warning for transmission lines, and in particular to a global safety early warning method and system based on wind vibration frequency monitoring for local transmission lines. Background Art

[0002] As the supporting body of UHV transmission lines, the transmission tower-line system is a complex coupling system with strong nonlinearity composed of conductors, insulators and transmission towers. Due to the high height and large span of the transmission tower-line system, the dynamic response under wind load is very complex and strong. Its damage and failure will lead to the paralysis of the entire power supply system, which will not only seriously affect people's production and living order, but also cause a variety of secondary disasters, causing very serious consequences to society and people's lives and property.

[0003] The existing wind vibration frequency monitoring system for transmission lines mainly installs multiple sensors on each transmission line segment on the transmission line to obtain multiple physical parameters during the operation of the transmission line, and then calculates the current wind vibration frequency of each transmission line segment, and performs status evaluation based on preset values ​​to achieve real-time wind vibration detection and early warning of the transmission line. This method requires the deployment of multiple sensors in each transmission line segment in advance. When the scale of power transmission is large, the cost and difficulty of deploying sensors are high. On the other hand, each sensor is only responsible for the transmission line segment it monitors, and the physical parameters of different segments are only used to evaluate their corresponding segments. It is impossible to achieve effective use of information and early warning. When the sensor of a segment fails, the monitoring system will completely lose control of the status information of the segment.

[0004] Another method is to realize wind-induced vibration monitoring and early warning of transmission lines by studying the response of a single transmission tower under wind-induced vibration. In dynamic analysis, the response of a single transmission tower under wind-induced vibration is much simpler than the response of the transmission tower-line system under wind-induced vibration after adding conductors. Therefore, wind-induced vibration tests at different wind directions can be conducted in advance in combination with the structure of the transmission tower itself and the terrain where it is located. The most unfavorable wind direction angle and wind-induced vibration coefficient for the transmission tower-line structure can be summarized, and on this basis, wind-induced vibration monitoring and early warning of transmission lines can be realized by monitoring the various physical parameters of the single transmission tower. This method is highly targeted and has high accuracy with the support of pre-experimental data. However, for large-scale and large-scale transmission lines, the terrains where different transmission towers are located are different, so this method is difficult to apply to large-scale and large-scale transmission scenarios. Summary of the invention

[0005] In response to the above technical problems, the present application provides a global safety warning method and system based on local transmission line wind vibration frequency monitoring, which realizes global safety warning of transmission lines by monitoring the wind vibration frequency at local locations in the transmission lines, thereby reducing deployment costs and improving the timeliness of warnings.

[0006] In a first aspect, the present application provides a global safety early warning method based on local transmission line wind vibration frequency monitoring, comprising:

[0007] Acquire real-time monitoring data corresponding to a plurality of preset positions of the transmission line, wherein the real-time monitoring data includes wind vibration frequency, wind speed and wind direction of the corresponding preset positions, and the preset positions are determined based on the importance and coupling connection degree of each position in the transmission line;

[0008] Inputting the multiple real-time monitoring data into a preset wind-vibration frequency prediction model, so that the wind-vibration frequency prediction model generates a prediction data set of wind-vibration frequency changes at various locations in the transmission line within the next preset time period according to the multiple real-time monitoring data and a preset transmission line topology map, wherein the wind-vibration frequency prediction model is constructed based on a multi-task learning model and obtained through training based on historical monitoring data;

[0009] Evaluate the predicted data set and a preset threshold value to obtain a safety level of each location in the transmission line;

[0010] Performing a comprehensive calculation based on the safety level of each position in the transmission line and the corresponding preset importance coefficient, to obtain the warning level of each position in the transmission line;

[0011] A safety warning is issued according to the warning level.

[0012] The embodiment of the present application provides a global safety warning method based on wind vibration frequency monitoring of local transmission lines. By acquiring real-time monitoring data of preset locations and inputting it into a wind vibration frequency prediction model for global wind vibration frequency prediction, a prediction data set of wind vibration frequency changes at various locations in the transmission line can be obtained. The safety level of each location is further determined based on the prediction data set, and the warning level of each location is obtained by comprehensively calculating the importance coefficient corresponding to each location. Finally, a safety warning is performed based on the warning level. The embodiment of the present application only needs to deploy sensors at designated locations to achieve global monitoring and early warning of the transmission line. Compared with the prior art, the sensor deployment cost and wind vibration monitoring cost are effectively reduced. In addition, the safety warning for the transmission line is based on the prediction data within the next preset time period, which can achieve early warning and improve the timeliness of the warning.

[0013] In a possible implementation, the wind vibration frequency prediction model generates a prediction data set of wind vibration frequency changes at various locations in the transmission line within the next preset time period according to the multiple real-time monitoring data and the preset transmission line topology map, including:

[0014] According to the location label of the real-time monitoring data, each real-time monitoring data is assigned to a corresponding predictor so that each predictor outputs a corresponding prediction result, wherein each predictor is obtained based on multi-task learning model training, and each prediction result includes a corresponding transmission line location, wind vibration transmission direction, and wind vibration transmission energy;

[0015] Perform several wind-vibration transmission simulations according to the prediction results, the wind-vibration frequencies and the transmission line topology diagram to generate the wind-vibration frequencies corresponding to each position in the transmission line at multiple moments in the future;

[0016] The wind vibration frequencies corresponding to various positions in the transmission line at multiple moments in the future are combined to construct a prediction data set of wind vibration frequency changes at various positions in the transmission line within the next preset time period.

[0017] The embodiment of the present application further illustrates the specific process of generating a prediction data set from a wind vibration frequency prediction model. First, prediction results for multiple specific locations can be obtained based on real-time monitoring data, and in the prediction process, each real-time monitoring data is assigned to a corresponding predictor to avoid using the same predictor for prediction, thereby improving the accuracy of the prediction. Furthermore, the characteristic that wind vibration can propagate in the transmission line is utilized. When wind vibration occurs at certain specific locations, the wind vibration will transmit the vibration to other locations of the transmission line based on the connection structure of the transmission line, thereby affecting other locations. Therefore, the embodiment of the present application performs multiple wind vibration transmission deductions based on the prediction results of multiple specific locations, obtains the wind vibration frequencies corresponding to each location in the transmission line at multiple times in the future and integrates them into a prediction data set, realizing the prediction of global information based on local information, providing data support for subsequent safety warnings, and effectively reducing the sensor deployment cost and wind vibration monitoring cost.

[0018] Furthermore, the wind vibration transmission simulation is performed several times according to each prediction result, the wind vibration frequency and the transmission line topology diagram to generate the wind vibration frequency corresponding to each position in the transmission line at multiple moments in the future, including:

[0019] The wind-induced vibration transfer simulation is performed based on the transmission line topology diagram, including input data, simulation simulation and prediction data, wherein in the first wind-induced vibration transfer simulation, the input data is the prediction results and the corresponding wind-induced vibration frequency;

[0020] In each wind vibration transmission simulation, the positions of each transmission line that will produce wind vibration frequency changes at the next moment are predicted based on the input data, and prediction data corresponding to each transmission line position is generated, and each prediction data is used as input data for the next wind vibration transmission simulation, wherein each prediction data includes the corresponding transmission line position, the wind vibration frequency at the corresponding position, the wind vibration transmission direction, and the wind vibration transmission energy;

[0021] When the number of wind vibration transfer simulations reaches a first preset value or the wind vibration transfer energy data obtained by the current wind vibration transfer simulation is less than a second preset value, the wind vibration transfer simulation is stopped to obtain the wind vibration frequencies corresponding to each position in the transmission line at multiple moments in the future.

[0022] The embodiment of the present application further illustrates the specific process of wind vibration transmission deduction. Since in the actual process of wind vibration transmission, wind vibration will transmit vibration to other positions of the transmission line based on the connection structure of the transmission line, the wind vibration transmission deduction of the embodiment of the present application is also based on the transmission line topology map, simulating the real process of wind vibration transmission as much as possible to improve the accuracy of the prediction. Among them, in the first wind vibration transmission deduction, relying on the prediction results output by each predictor and the wind vibration frequency monitored in real time, the positions of each transmission line that will produce wind vibration frequency changes at the next moment and the corresponding prediction data can be deduced based on these data. And each subsequent wind vibration transmission deduction can be deduced based on the prediction data at the current moment to obtain the prediction data at the next moment. Through multiple wind vibration transmission deductions, the global wind vibration frequency data of the transmission line at each moment within the preset time period can be finally obtained. In addition, considering the combined influence of different wind vibration frequencies and different wind vibration transmission directions at each position on a certain position, the embodiment of the present application uses the prediction data of multiple positions for comprehensive judgment in each wind vibration transmission deduction, which ensures the accuracy of the deduction prediction.

[0023] In a possible implementation, the wind vibration frequency prediction model is constructed based on a multi-task learning model and trained based on historical monitoring data, including:

[0024] Dividing each location in the transmission line into T areas according to a preset number of tasks T;

[0025] Based on the multi-task learning model and the support vector machine model, an initial wind-induced vibration frequency prediction model is constructed, wherein the initial wind-induced vibration frequency prediction model includes T initial predictors corresponding to each region, each of the initial predictors has shared parameters, and each of the initial predictors has independent parameters.

[0026] The T initial predictors are trained simultaneously by using the historical monitoring data, the shared parameters and each independent parameter are updated, and thus the wind vibration frequency prediction model is obtained.

[0027] The embodiment of the present application provides a training method for a wind-induced vibration frequency prediction model. First, the transmission line is divided into several areas, each area corresponds to a learning task. Since there is a strong correlation between the various learning tasks, the embodiment of the present application uses a multi-task learning method to train the initial wind-induced vibration frequency prediction model. By setting shared parameters, the model can learn multiple tasks at the same time. Each task can use the knowledge learned from other tasks to assist its own learning, thereby improving the training efficiency of the model. In the subsequent prediction process, different prediction tasks can be targetedly assigned to different predictors to improve the prediction accuracy of the model.

[0028] Furthermore, the initial wind vibration frequency prediction model is constructed based on the multi-task learning model and the support vector machine model. The specific formula is:

[0029] f t =w ′ t x+b t

[0030] w t =w0+v t

[0031] Among them, f t is the tth initial predictor, w t and b t is the parameter to be determined for the tth initial predictor, w0 is the shared parameter of each initial predictor, v t are the independent parameters of the tth initial predictor.

[0032] In the second aspect, accordingly, the present application provides a global safety warning system based on local transmission line wind vibration frequency monitoring, including an acquisition module, a prediction module, an evaluation module, a comprehensive calculation module and a warning module;

[0033] The acquisition module is used to acquire real-time monitoring data corresponding to a plurality of preset positions of the transmission line, wherein the real-time monitoring data includes wind vibration frequency, wind speed and wind direction of the corresponding preset positions, and the preset positions are determined based on the importance and coupling connection degree of each position in the transmission line;

[0034] The prediction module is used to input the multiple real-time monitoring data into a preset wind-vibration frequency prediction model, so that the wind-vibration frequency prediction model generates a prediction data set of wind-vibration frequency changes at various locations in the transmission line within the next preset time period according to the multiple real-time monitoring data and a preset transmission line topology map, wherein the wind-vibration frequency prediction model is constructed based on a multi-task learning model and obtained through training based on historical monitoring data;

[0035] The evaluation module is used to evaluate according to the prediction data set and the preset threshold value to obtain the safety level of each position in the transmission line;

[0036] The comprehensive calculation module is used to perform comprehensive calculation according to the safety level of each position in the transmission line and the corresponding preset importance coefficient, so as to obtain the warning level of each position in the transmission line;

[0037] The warning module is used to issue a safety warning according to the warning level.

[0038] In a possible implementation, the prediction module includes a prediction unit, a deduction unit, and a data set construction unit;

[0039] The prediction unit is used to assign each real-time monitoring data to a corresponding predictor according to the location label of the real-time monitoring data so that each predictor outputs a corresponding prediction result, wherein each predictor is obtained based on multi-task learning model training, and each prediction result includes a corresponding transmission line location, wind vibration transmission direction, and wind vibration transmission energy;

[0040] The deduction unit is used to perform several wind-vibration transmission deductions according to various prediction results, wind-vibration frequencies and the transmission line topology diagram, and generate wind-vibration frequencies corresponding to various positions in the transmission line at multiple moments in the future;

[0041] The data set construction unit is used to combine the wind vibration frequencies corresponding to each position in the transmission line at multiple moments in the future to construct a prediction data set of the wind vibration frequency changes at each position in the transmission line within the next preset time period.

[0042] Furthermore, the deduction unit performs several wind-induced vibration transmission deductions according to the prediction results, the wind-induced vibration frequencies and the transmission line topology diagram, and generates wind-induced vibration frequencies corresponding to various positions in the transmission line at multiple moments in the future, including:

[0043] The wind-induced vibration transfer simulation is performed based on the transmission line topology diagram, including input data, simulation simulation and prediction data, wherein in the first wind-induced vibration transfer simulation, the input data is the prediction results and the corresponding wind-induced vibration frequency;

[0044] In each wind vibration transmission simulation, the positions of each transmission line that will produce wind vibration frequency changes at the next moment are predicted based on the input data, and prediction data corresponding to each transmission line position is generated, and each prediction data is used as input data for the next wind vibration transmission simulation, wherein each prediction data includes the corresponding transmission line position, the wind vibration frequency at the corresponding position, the wind vibration transmission direction, and the wind vibration transmission energy;

[0045] When the number of wind vibration transfer simulations reaches a first preset value or the wind vibration transfer energy data obtained by the current wind vibration transfer simulation is less than a second preset value, the wind vibration transfer simulation is stopped to obtain the wind vibration frequencies corresponding to each position in the transmission line at multiple moments in the future.

[0046] In a possible implementation, the global safety warning system further includes a model building module, which is used to build the wind vibration frequency prediction model based on a multi-task learning model and train based on historical monitoring data, including a partitioning unit, an initial model building unit, and a model training unit;

[0047] Wherein, the partition unit is used to divide each position in the transmission line into T areas according to a preset number of tasks T;

[0048] The initial model building unit is used to build an initial wind-induced vibration frequency prediction model based on a multi-task learning model and a support vector machine model, wherein the initial wind-induced vibration frequency prediction model includes T initial predictors corresponding to each region, each of the initial predictors has shared parameters, and each of the initial predictors has independent parameters.

[0049] The model training unit is used to simultaneously train the T initial predictors through the historical monitoring data, update the shared parameters and each independent parameter, and thus obtain the wind vibration frequency prediction model.

[0050] Furthermore, the initial model building unit builds an initial wind vibration frequency prediction model based on the multi-task learning model and the support vector machine model. The specific formula is:

[0051] f t =w ′ t x+b t

[0052] w t =w0+v t

[0053] Among them, f t is the tth initial predictor, w t and b t is the parameter to be determined for the tth initial predictor, w0 is the shared parameter of each initial predictor, v t are the independent parameters of the tth initial predictor. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 : A flow chart of a global safety warning method based on local transmission line wind vibration frequency monitoring provided in an embodiment of the present application.

[0055] Figure 2 : A flow chart of predicting global wind-induced vibration frequency in a global safety warning method based on local transmission line wind-induced vibration frequency monitoring provided in an embodiment of the present application.

[0056] Figure 3 : A structural schematic diagram of a global safety warning system based on local transmission line wind vibration frequency monitoring provided in an embodiment of the present application.

[0057] Figure 4 : A structural schematic diagram of a prediction module of a global safety warning system based on local transmission line wind vibration frequency monitoring provided in an embodiment of the present application.

[0058] Figure 5 : A structural diagram of a model building module of a global safety warning system based on local transmission line wind vibration frequency monitoring provided in an embodiment of the present application. DETAILED DESCRIPTION

[0059] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0060] It should be noted that the step numbers in the text are only for the convenience of explanation of the specific embodiments and do not serve to limit the order in which the steps are executed. In the description of this application, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features.

[0061] Throughout the specification, the multi-task learning described in this specification is a learning method in the field of machine learning technology, and its goal is to use the useful information contained in multiple learning tasks to help learn a more accurate learner for each task. When the various learning tasks are related tasks, joint learning of multiple tasks can achieve better performance than learning them separately. Multi-task learning can be seen as a way for machines to imitate human learning behavior, because humans often transfer the knowledge of one task to another related task. For example, the skills of playing badminton and playing tennis can help improve each other. Similar to human learning, it is useful for the model to learn multiple learning tasks at the same time, because one task can utilize the knowledge of another related task.

[0062] Embodiment 1:

[0063] like Figure 1 As shown, embodiment 1 provides a global safety early warning method based on local transmission line wind vibration frequency monitoring, including steps S1-S5:

[0064] Step S1, obtaining real-time monitoring data corresponding to a plurality of preset positions of the transmission line, wherein the real-time monitoring data includes wind vibration frequency, wind speed and wind direction of the corresponding preset positions, and the preset positions are determined based on the importance and coupling connection degree of each position in the transmission line;

[0065] Step S2, inputting the multiple real-time monitoring data into a preset wind-vibration frequency prediction model, so that the wind-vibration frequency prediction model generates a prediction data set of wind-vibration frequency changes at various locations in the transmission line within the next preset time period according to the multiple real-time monitoring data and a preset transmission line topology map, wherein the wind-vibration frequency prediction model is constructed based on a multi-task learning model and obtained through training based on historical monitoring data;

[0066] Step S3, evaluating the predicted data set and a preset threshold to obtain a safety level of each location in the transmission line;

[0067] Step S4, performing comprehensive calculation according to the safety level of each position in the transmission line and the corresponding preset importance coefficient, to obtain the warning level of each position in the transmission line;

[0068] Step S5: issuing a safety warning according to the warning level.

[0069] The embodiment of the present application provides a global safety warning method based on wind vibration frequency monitoring of local transmission lines. By acquiring real-time monitoring data of preset locations and inputting it into a wind vibration frequency prediction model for global wind vibration frequency prediction, a prediction data set of wind vibration frequency changes at various locations in the transmission line can be obtained. The safety level of each location is further determined based on the prediction data set, and the warning level of each location is obtained by comprehensively calculating the importance coefficient corresponding to each location. Finally, a safety warning is performed based on the warning level. The embodiment of the present application only needs to deploy sensors at designated locations to achieve global monitoring and early warning of the transmission line. Compared with the prior art, the sensor deployment cost and wind vibration monitoring cost are effectively reduced. In addition, the safety warning for the transmission line is based on the prediction data within the next preset time period, which can achieve early warning and improve the timeliness of the warning.

[0070] In a possible implementation, in step S2, the wind vibration frequency prediction model generates a prediction data set of wind vibration frequency changes at various locations in the transmission line within the next preset time period based on the multiple real-time monitoring data and the preset transmission line topology map, such as Figure 2 As shown, steps S201-S203 are included:

[0071] Step S201: according to the location label of the real-time monitoring data, each real-time monitoring data is assigned to a corresponding predictor so that each predictor outputs a corresponding prediction result, wherein each predictor is obtained based on multi-task learning model training, and each prediction result includes a corresponding transmission line location, wind vibration transmission direction, and wind vibration transmission energy;

[0072] Step S202: performing several wind-induced vibration transmission simulations according to the prediction results, the wind-induced vibration frequencies and the transmission line topology diagram, and generating the wind-induced vibration frequencies corresponding to each position in the transmission line at multiple moments in the future;

[0073] Step S203: combining the wind vibration frequencies corresponding to the various positions in the transmission line at multiple moments in the future, and constructing a prediction data set of the wind vibration frequency changes at the various positions in the transmission line in the next preset time period.

[0074] The embodiment of the present application further illustrates the specific process of generating a prediction data set from a wind vibration frequency prediction model. First, prediction results for multiple specific locations can be obtained based on real-time monitoring data, and in the prediction process, each real-time monitoring data is assigned to a corresponding predictor to avoid using the same predictor for prediction, thereby improving the accuracy of the prediction. Furthermore, the characteristic that wind vibration can propagate in the transmission line is utilized. When wind vibration occurs at certain specific locations, the wind vibration will transmit the vibration to other locations of the transmission line based on the connection structure of the transmission line, thereby affecting other locations. Therefore, the embodiment of the present application performs multiple wind vibration transmission deductions based on the prediction results of multiple specific locations, obtains the wind vibration frequencies corresponding to each location in the transmission line at multiple times in the future and integrates them into a prediction data set, realizing the prediction of global information based on local information, providing data support for subsequent safety warnings, and effectively reducing the sensor deployment cost and wind vibration monitoring cost.

[0075] Further, in step S202, the wind vibration transmission simulation is performed several times according to each prediction result, the wind vibration frequency and the transmission line topology diagram to generate the wind vibration frequency corresponding to each position in the transmission line at multiple moments in the future, including:

[0076] The wind-induced vibration transfer simulation is performed based on the transmission line topology diagram, including input data, simulation simulation and prediction data, wherein in the first wind-induced vibration transfer simulation, the input data is the prediction results and the corresponding wind-induced vibration frequency;

[0077] In each wind vibration transmission simulation, the positions of each transmission line that will produce wind vibration frequency changes at the next moment are predicted based on the input data, and prediction data corresponding to each transmission line position is generated, and each prediction data is used as input data for the next wind vibration transmission simulation, wherein each prediction data includes the corresponding transmission line position, the wind vibration frequency at the corresponding position, the wind vibration transmission direction, and the wind vibration transmission energy;

[0078] When the number of wind vibration transfer simulations reaches a first preset value or the wind vibration transfer energy data obtained by the current wind vibration transfer simulation is less than a second preset value, the wind vibration transfer simulation is stopped to obtain the wind vibration frequencies corresponding to each position in the transmission line at multiple moments in the future.

[0079] The embodiment of the present application further illustrates the specific process of wind vibration transmission deduction. Since in the actual process of wind vibration transmission, wind vibration will transmit vibration to other positions of the transmission line based on the connection structure of the transmission line, the wind vibration transmission deduction of the embodiment of the present application is also based on the transmission line topology map, simulating the real process of wind vibration transmission as much as possible to improve the accuracy of the prediction. Among them, in the first wind vibration transmission deduction, relying on the prediction results output by each predictor and the wind vibration frequency monitored in real time, the positions of each transmission line that will produce wind vibration frequency changes at the next moment and the corresponding prediction data can be deduced based on these data. And each subsequent wind vibration transmission deduction can be deduced based on the prediction data at the current moment to obtain the prediction data at the next moment. Through multiple wind vibration transmission deductions, the global wind vibration frequency data of the transmission line at each moment within the preset time period can be finally obtained. In addition, considering the combined influence of different wind vibration frequencies and different wind vibration transmission directions at each position on a certain position, the embodiment of the present application uses the prediction data of multiple positions for comprehensive judgment in each wind vibration transmission deduction, which ensures the accuracy of the deduction prediction.

[0080] In a possible implementation, the wind vibration frequency prediction model is constructed based on a multi-task learning model and trained based on historical monitoring data, including:

[0081] Dividing each location in the transmission line into T areas according to a preset number of tasks T;

[0082] Based on the multi-task learning model and the support vector machine model, an initial wind-induced vibration frequency prediction model is constructed, wherein the initial wind-induced vibration frequency prediction model includes T initial predictors corresponding to each region, each of the initial predictors has shared parameters, and each of the initial predictors has independent parameters.

[0083] The T initial predictors are trained simultaneously by using the historical monitoring data, the shared parameters and each independent parameter are updated, and thus the wind vibration frequency prediction model is obtained.

[0084] The embodiment of the present application provides a training method for a wind-induced vibration frequency prediction model. First, the transmission line is divided into several areas, each area corresponds to a learning task. Since there is a strong correlation between the various learning tasks, the embodiment of the present application uses a multi-task learning method to train the initial wind-induced vibration frequency prediction model. By setting shared parameters, the model can learn multiple tasks at the same time. Each task can use the knowledge learned from other tasks to assist its own learning, thereby improving the training efficiency of the model. In the subsequent prediction process, different prediction tasks can be targetedly assigned to different predictors to improve the prediction accuracy of the model.

[0085] Furthermore, the initial wind vibration frequency prediction model is constructed based on the multi-task learning model and the support vector machine model. The specific formula is:

[0086] f t =w ′ t x+b t

[0087] w t =w0+v t

[0088] Among them, f t is the tth initial predictor, w t and b t is the parameter to be determined for the tth initial predictor, w0 is the shared parameter of each initial predictor, v t are the independent parameters of the tth initial predictor.

[0089] The T initial predictors are trained simultaneously by the historical monitoring data, the shared parameters and each independent parameter are updated, and then the wind vibration frequency prediction model is obtained, specifically:

[0090] The objective function and corresponding constraints of the wind vibration frequency prediction model are constructed as follows:

[0091]

[0092] Subject to the following constraints:

[0093] w ′ t x ti +b t -y ti ≤ε ti +ξ ti

[0094]

[0095] Among them, T is the total number of tasks, λ1 and λ2 are the t and w0 are regularization constraints imposed to control the correlation between tasks. N is the total amount of historical monitoring data. C t is the penalty parameter used to balance the margin and training error of each predictor, ε ti is the tolerable deviation of each sample, ξ ti and is a slack variable used to measure the extent to which each sample exceeds the tolerable deviation, y ti is the true label of each piece of historical monitoring data.

[0096] Furthermore, the objective function can be dually optimized by Lagrange multiplier method, and after inputting historical monitoring data, the determined values ​​of shared parameters and each independent parameter can be solved to obtain the wind vibration frequency prediction model.

[0097] In the second aspect, accordingly, the present application provides a global safety warning system based on local transmission line wind vibration frequency monitoring, such as Figure 3 As shown, it includes an acquisition module 10, a prediction module 20, an evaluation module 30, a comprehensive calculation module 40 and an early warning module 50;

[0098] The acquisition module 10 is used to acquire real-time monitoring data corresponding to a plurality of preset positions of the transmission line, wherein the real-time monitoring data includes wind vibration frequency, wind speed and wind direction of the corresponding preset positions, and the preset positions are determined based on the importance and coupling connection degree of each position in the transmission line;

[0099] The prediction module 20 is used to input the multiple real-time monitoring data into a preset wind vibration frequency prediction model, so that the wind vibration frequency prediction model generates a prediction data set of wind vibration frequency changes at various locations in the transmission line within the next preset time period according to the multiple real-time monitoring data and a preset transmission line topology map, wherein the wind vibration frequency prediction model is constructed based on a multi-task learning model and obtained through training based on historical monitoring data;

[0100] The evaluation module 30 is used to evaluate according to the prediction data set and the preset threshold value to obtain the safety level of each location in the transmission line;

[0101] The comprehensive calculation module 40 is used to perform comprehensive calculation according to the safety level of each position in the transmission line and the corresponding preset importance coefficient to obtain the warning level of each position in the transmission line;

[0102] The warning module 50 is used to issue a safety warning according to the warning level.

[0103] In one possible implementation, Figure 4 As shown, the prediction module 20 includes a prediction unit 201, a deduction unit 202 and a data set construction unit 203;

[0104] The prediction unit 201 is used to assign each real-time monitoring data to a corresponding predictor according to the location label of the real-time monitoring data so that each predictor outputs a corresponding prediction result, wherein each predictor is obtained based on multi-task learning model training, and each prediction result includes a corresponding transmission line location, wind vibration transmission direction, and wind vibration transmission energy;

[0105] The deduction unit 202 is used to perform several wind-induced vibration transmission deductions according to various prediction results, wind-induced vibration frequencies and the transmission line topology diagram, and generate wind-induced vibration frequencies corresponding to various positions in the transmission line at multiple moments in the future;

[0106] The data set construction unit 203 is used to combine the wind vibration frequencies corresponding to each position in the transmission line at multiple moments in the future to construct a prediction data set of wind vibration frequency changes at each position in the transmission line in the next preset time period.

[0107] Furthermore, the deduction unit 202 performs several wind-induced vibration transmission deductions according to the prediction results, the wind-induced vibration frequencies and the transmission line topology diagram, and generates the wind-induced vibration frequencies corresponding to each position in the transmission line at multiple moments in the future, including:

[0108] The wind-induced vibration transfer simulation is performed based on the transmission line topology diagram, including input data, simulation simulation and prediction data, wherein in the first wind-induced vibration transfer simulation, the input data is the prediction results and the corresponding wind-induced vibration frequency;

[0109] In each wind vibration transmission simulation, the positions of each transmission line that will produce wind vibration frequency changes at the next moment are predicted based on the input data, and prediction data corresponding to each transmission line position is generated, and each prediction data is used as input data for the next wind vibration transmission simulation, wherein each prediction data includes the corresponding transmission line position, the wind vibration frequency at the corresponding position, the wind vibration transmission direction, and the wind vibration transmission energy;

[0110] When the number of wind vibration transfer simulations reaches a first preset value or the wind vibration transfer energy data obtained by the current wind vibration transfer simulation is less than a second preset value, the wind vibration transfer simulation is stopped to obtain the wind vibration frequencies corresponding to each position in the transmission line at multiple moments in the future.

[0111] In one possible implementation, Figure 5As shown, the global safety warning system further includes a model building module 60, which is used to build the wind vibration frequency prediction model based on a multi-task learning model and train based on historical monitoring data, including a partitioning unit 601, an initial model building unit 602 and a model training unit 603;

[0112] The partition unit 601 is used to divide each position in the power transmission line into T areas according to a preset number of tasks T;

[0113] The initial model building unit 602 is used to build an initial wind-induced vibration frequency prediction model based on a multi-task learning model and a support vector machine model, wherein the initial wind-induced vibration frequency prediction model includes T initial predictors corresponding to each region, each of the initial predictors has shared parameters, and each of the initial predictors has independent parameters.

[0114] The model training unit 603 is used to simultaneously train the T initial predictors through the historical monitoring data, update the shared parameters and each independent parameter, and then obtain the wind vibration frequency prediction model.

[0115] Furthermore, the initial model building unit 602 builds an initial wind vibration frequency prediction model based on the multi-task learning model and the support vector machine model. The specific formula is:

[0116] f t =w ′ t x+b t

[0117] w t =w0+v t

[0118] Among them, f t is the tth initial predictor, w t and b t is the parameter to be determined for the tth initial predictor, w0 is the shared parameter of each initial predictor, v t are the independent parameters of the tth initial predictor.

[0119] The more detailed working principle and step flow of this embodiment can refer to, but are not limited to, the relevant records of Embodiment 1.

[0120] The specific embodiments described above further describe the purpose, technical solutions and beneficial effects of the present application in detail. It should be understood that the above description is only a specific embodiment of the present application and is not intended to limit the scope of protection of the present application. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of protection of the present application.

Claims

1. A global safety early warning method based on local transmission line wind vibration frequency monitoring, characterized in that: include: Acquire real-time monitoring data corresponding to a plurality of preset positions of the transmission line, wherein the real-time monitoring data includes wind vibration frequency, wind speed and wind direction of the corresponding preset positions, and the preset positions are determined based on the importance and coupling connection degree of each position in the transmission line; Inputting the multiple real-time monitoring data into a preset wind-vibration frequency prediction model, so that the wind-vibration frequency prediction model generates a prediction data set of wind-vibration frequency changes at various locations in the transmission line within the next preset time period according to the multiple real-time monitoring data and a preset transmission line topology map, wherein the wind-vibration frequency prediction model is constructed based on a multi-task learning model and obtained through training based on historical monitoring data; Evaluate the predicted data set and a preset threshold value to obtain a safety level of each location in the transmission line; Performing a comprehensive calculation based on the safety level of each position in the transmission line and the corresponding preset importance coefficient, to obtain the warning level of each position in the transmission line; A safety warning is issued according to the warning level.

2. A global safety early warning method based on local transmission line wind vibration frequency monitoring as claimed in claim 1, characterized in that: The wind vibration frequency prediction model generates a prediction data set of wind vibration frequency changes at various locations in the transmission line within the next preset time period according to the multiple real-time monitoring data and the preset transmission line topology map, including: According to the location label of the real-time monitoring data, each real-time monitoring data is assigned to a corresponding predictor so that each predictor outputs a corresponding prediction result, wherein each predictor is obtained based on multi-task learning model training, and each prediction result includes a corresponding transmission line location, wind vibration transmission direction, and wind vibration transmission energy; Perform several wind-vibration transmission simulations according to the prediction results, the wind-vibration frequencies and the transmission line topology diagram to generate the wind-vibration frequencies corresponding to each position in the transmission line at multiple moments in the future; The wind vibration frequencies corresponding to various positions in the transmission line at multiple moments in the future are combined to construct a prediction data set of wind vibration frequency changes at various positions in the transmission line within the next preset time period.

3. A global safety early warning method based on local transmission line wind vibration frequency monitoring as claimed in claim 2, characterized in that: The method of performing several wind-induced vibration transmission simulations according to the prediction results, the wind-induced vibration frequencies and the transmission line topology diagram to generate the wind-induced vibration frequencies corresponding to the various positions in the transmission line at multiple moments in the future includes: The wind-induced vibration transfer simulation is performed based on the transmission line topology diagram, including input data, simulation simulation and prediction data, wherein in the first wind-induced vibration transfer simulation, the input data is the prediction results and the corresponding wind-induced vibration frequency; In each wind vibration transmission simulation, the positions of each transmission line that will produce wind vibration frequency changes at the next moment are predicted based on the input data, and prediction data corresponding to each transmission line position is generated, and each prediction data is used as input data for the next wind vibration transmission simulation, wherein each prediction data includes the corresponding transmission line position, the wind vibration frequency at the corresponding position, the wind vibration transmission direction, and the wind vibration transmission energy; When the number of wind vibration transfer simulations reaches a first preset value or the wind vibration transfer energy data obtained by the current wind vibration transfer simulation is less than a second preset value, the wind vibration transfer simulation is stopped to obtain the wind vibration frequencies corresponding to each position in the transmission line at multiple moments in the future.

4. A global safety early warning method based on local transmission line wind vibration frequency monitoring as claimed in claim 1, characterized in that: The wind vibration frequency prediction model is constructed based on a multi-task learning model and trained based on historical monitoring data, including: Dividing each location in the transmission line into T areas according to a preset number of tasks T; Based on the multi-task learning model and the support vector machine model, an initial wind-induced vibration frequency prediction model is constructed, wherein the initial wind-induced vibration frequency prediction model includes T initial predictors corresponding to each region, each of the initial predictors has shared parameters, and each of the initial predictors has independent parameters. The T initial predictors are trained simultaneously by using the historical monitoring data, the shared parameters and each independent parameter are updated, and thus the wind vibration frequency prediction model is obtained.

5. A global safety early warning method based on local transmission line wind vibration frequency monitoring as claimed in claim 4, characterized in that: The initial wind vibration frequency prediction model is constructed based on the multi-task learning model and the support vector machine model. The specific formula is: f t =w ′ t x+b t In t =w0+v t Among them, f t is the tth initial predictor, w t and b t is the parameter to be determined for the tth initial predictor, w0 is the shared parameter of each initial predictor, v t are the independent parameters of the tth initial predictor.

6. A global safety warning system based on local transmission line wind vibration frequency monitoring, characterized in that: It includes acquisition module, prediction module, evaluation module, comprehensive calculation module and early warning module; The acquisition module is used to acquire real-time monitoring data corresponding to a plurality of preset positions of the transmission line, wherein the real-time monitoring data includes wind vibration frequency, wind speed and wind direction of the corresponding preset positions, and the preset positions are determined based on the importance and coupling connection degree of each position in the transmission line; The prediction module is used to input the multiple real-time monitoring data into a preset wind-vibration frequency prediction model, so that the wind-vibration frequency prediction model generates a prediction data set of wind-vibration frequency changes at various locations in the transmission line within the next preset time period according to the multiple real-time monitoring data and a preset transmission line topology map, wherein the wind-vibration frequency prediction model is constructed based on a multi-task learning model and obtained through training based on historical monitoring data; The evaluation module is used to evaluate according to the prediction data set and the preset threshold value to obtain the safety level of each position in the transmission line; The comprehensive calculation module is used to perform comprehensive calculation according to the safety level of each position in the transmission line and the corresponding preset importance coefficient, so as to obtain the warning level of each position in the transmission line; The warning module is used to issue a safety warning according to the warning level.

7. A global safety warning system based on local transmission line wind vibration frequency monitoring as claimed in claim 6, characterized in that: The prediction module includes a prediction unit, a deduction unit and a data set construction unit; The prediction unit is used to assign each real-time monitoring data to a corresponding predictor according to the location label of the real-time monitoring data so that each predictor outputs a corresponding prediction result, wherein each predictor is obtained based on multi-task learning model training, and each prediction result includes a corresponding transmission line location, wind vibration transmission direction, and wind vibration transmission energy; The deduction unit is used to perform several wind-vibration transmission deductions according to various prediction results, wind-vibration frequencies and the transmission line topology diagram, and generate wind-vibration frequencies corresponding to various positions in the transmission line at multiple moments in the future; The data set construction unit is used to combine the wind vibration frequencies corresponding to each position in the transmission line at multiple moments in the future to construct a prediction data set of the wind vibration frequency changes at each position in the transmission line within the next preset time period.

8. A global safety warning system based on wind vibration frequency monitoring of local transmission lines as claimed in claim 7, characterized in that: The deduction unit performs several wind-induced vibration transmission deductions according to the prediction results, the wind-induced vibration frequencies and the transmission line topology diagram, and generates wind-induced vibration frequencies corresponding to various positions in the transmission line at multiple moments in the future, including: The wind-induced vibration transfer simulation is performed based on the transmission line topology diagram, including input data, simulation simulation and prediction data, wherein in the first wind-induced vibration transfer simulation, the input data is the prediction results and the corresponding wind-induced vibration frequency; In each wind vibration transmission simulation, the positions of each transmission line that will produce wind vibration frequency changes at the next moment are predicted based on the input data, and prediction data corresponding to each transmission line position is generated, and each prediction data is used as input data for the next wind vibration transmission simulation, wherein each prediction data includes the corresponding transmission line position, the wind vibration frequency at the corresponding position, the wind vibration transmission direction, and the wind vibration transmission energy; When the number of wind vibration transfer simulations reaches a first preset value or the wind vibration transfer energy data obtained by the current wind vibration transfer simulation is less than a second preset value, the wind vibration transfer simulation is stopped to obtain the wind vibration frequencies corresponding to each position in the transmission line at multiple moments in the future.

9. A global safety early warning system based on local transmission line wind vibration frequency monitoring as claimed in claim 6, characterized in that: The global safety warning system also includes a model building module, which is used to build the wind vibration frequency prediction model based on a multi-task learning model and train based on historical monitoring data, including a partitioning unit, an initial model building unit and a model training unit; Wherein, the partition unit is used to divide each position in the transmission line into T areas according to a preset number of tasks T; The initial model building unit is used to build an initial wind-induced vibration frequency prediction model based on a multi-task learning model and a support vector machine model, wherein the initial wind-induced vibration frequency prediction model includes T initial predictors corresponding to each region, each of the initial predictors has shared parameters, and each of the initial predictors has independent parameters. The model training unit is used to simultaneously train the T initial predictors through the historical monitoring data, update the shared parameters and each independent parameter, and thus obtain the wind vibration frequency prediction model.

10. A global safety warning system based on local transmission line wind vibration frequency monitoring as claimed in claim 9, characterized in that: The initial model building unit builds an initial wind vibration frequency prediction model based on a multi-task learning model and a support vector machine model. The specific formula is: f t =w ′ t x+b t In t =w0+v t Among them, f t is the tth initial predictor, w t and b t is the parameter to be determined for the tth initial predictor, w0 is the shared parameter of each initial predictor, v t are the independent parameters of the tth initial predictor.