Power transmission line icing early warning method based on weather factors

Abstract

The invention discloses a power transmission line icing early warning method based on weather factors, and belongs to the technical field of power transmission line inspection. According to the invention, the problem of insufficient advanced time of an existing power transmission line icing early warning method is solved. According to the method, an icing susceptibility degree description acquisition module and an ultra-long advance periodic icing feature calculation module are constructed, typical weather cycle conditions can be described by using the two modules, then an ultra-long advance period identification model is obtained, and then ultra-long time advance early warning is carried out in power transmission line icing high-incidence areas. By means of the method, it can be found that the weather cycle period in which the weather is likely to enter the icing high-incidence weather, the indirect incidence relation between the weather and icing is established, and the weather state in which the weather state related to icing can be entered with the large probability under the continuous multi-day condition is found; and the icing condition can be pre-warned for many days in advance. The method can be applied to power transmission line icing early warning.

Description

Technical Field

[0001] The present invention belongs to the technical field of power transmission line inspection, and in particular relates to a transmission line icing early warning method based on weather factors. Background Art

[0002] In some remote areas in northern my country (such as valleys, river canyons, etc.), transmission lines are prone to icing. Once the transmission lines are covered with ice, it takes a long time to prepare for de-icing, such as checking the automatic de-icing equipment, powering on in advance, and the de-icing team heading to the transmission line iced area. However, in the process of de-icing preparation, the icing of the transmission lines will have an impact on power transmission. Therefore, it is very necessary to carry out ultra-long advance icing warnings in areas with high incidence of transmission line icing, so as to make timely and effective preparations for icing and reduce the disaster losses that may be caused by icing of the transmission lines.

[0003] The technical means used in the existing transmission line icing warning methods mainly include: 1. Based on the detailed daily temperature, humidity, rainfall and snowfall conditions, combined with the data collected by the regional cameras, an ice warning is issued when the relevant indicators exceed a certain value and last for a period of time. However, this ice warning method has a short lead time. When the warning is issued, the relevant transmission lines are basically in the process of ice accumulation, and countermeasures need to be taken immediately; 2. Carry out forecasts based on micro-scale weather conditions, conduct a joint analysis of certain weather characteristics and icing conditions, and find the weather conditions that directly lead to icing. This type of technology can provide icing warnings about one day in advance, and relevant inspection units can preheat de-icing equipment in advance and make preliminary preparations for de-icing.

[0004] The above two types of transmission line icing warning technologies are more effective in dealing with small areas, urban areas and low-incidence icing. However, for some remote cols, river canyons, leeward slopes and other areas in the northern region, these areas are prone to icing. In the event of icing, a long advance time is required to preheat the deicing equipment and the deicing team needs to enter the icing area. At this time, the advance time of the two existing transmission line icing warning methods is not enough to meet the requirements. In the case of insufficient advance time, the relevant inspection units may not respond adequately when icing occurs, resulting in serious accidents such as the collapse of transmission towers. On the other hand, excessive response to icing leads to excessive loss of electricity, equipment and human resources, and great management costs. Therefore, it is very necessary to conduct ultra-long advance warnings for areas with high icing incidence. Summary of the invention

[0005] The purpose of the present invention is to solve the problem of insufficient advance time of the existing transmission line icing warning method, and propose a transmission line icing warning method based on weather factors.

[0006] The technical solution adopted by the present invention to solve the above technical problems is: a transmission line icing early warning method based on weather factors, the method specifically comprising the following steps: Step S1, inputting a list of historical data of weather and icing occurrence in a region History, the number of early warning days YJLengh and the length of the period to be identified ZQLength; According to the historical data list History, the mean temperature JZWD, mean humidity JZSD, mean wind speed JZFS and snowfall correlation JZJX in the statistical area under ice cover conditions; Step S2: Establish an icing susceptibility description acquisition module MYFVector. The input of the module MYFVector is an icing susceptibility description input variable MYFInput with the same structure as each record in the history data list History. The output is an icing susceptibility description output vector ; Step S3, constructing an advance periodic icing feature calculation module MCCFeature, the input of MCCFeature is a periodic list MCCFeatureInput consisting of records of consecutive ZQLength days, and the structure of each record in the periodic list MCCFeatureInput is the same as the record structure of the historical data list History; The module MCCFeature uses MYFVector to process each record in the periodic list MCCFeatureInput separately to obtain the advance periodic icing feature output MCCFeatureOutput; Step S4, constructing an advance periodic icing feature table TZTable using MCCFeature and History, and obtaining an advance periodic identification model SModel based on TZTable; Step S5, input the weather data BTest of the recent ZQLength days in the area to be warned of transmission line icing disaster, and use SModel to warn whether transmission line icing will occur in the area to be warned of transmission line icing disaster in YJLengh days.

[0007] The beneficial effects of the present invention are: The present invention provides a method for super-long advance warning of areas with high incidence of icing on transmission lines based on the identification of typical weather cycles. First, an icing susceptibility description acquisition module and a super-long advance periodic icing feature calculation module are constructed. These two modules can be used to describe typical weather cycle conditions and then obtain a super-long advance period identification model, and then perform super-long advance warning of areas with high incidence of icing on transmission lines. The method of the present invention can be used to find the weather cycle that is bound to enter high-incidence weather of icing, establish an indirect correlation between weather and icing, and find the weather state associated with icing with a high probability when it lasts for many days; the icing situation can be warned many days in advance, so that the inspection unit can make timely and effective preparations for dealing with icing, reduce the disaster losses that may be caused by icing, and prevent excessive protection to effectively reduce the management cost of maintenance work. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 The present invention is a flow chart of a transmission line icing warning method based on weather factors. DETAILED DESCRIPTION

[0009] Specific implementation method 1: Combination Figure 1 The present embodiment describes a method for early warning of ice covering of transmission lines based on weather factors, and the method specifically comprises the following steps: Step S1, inputting a list of historical data of weather and icing occurrence in a region (here, the region is a high-incidence region of icing disasters), the number of early warning days YJLengh, and the length of the period to be identified ZQLength; According to the historical data list History, the mean temperature JZWD, mean humidity JZSD, mean wind speed JZFS and snowfall correlation JZJX in the statistical area under ice cover conditions; Step S2: Establish an icing susceptibility description acquisition module MYFVector. The input of the module MYFVector is an icing susceptibility description input variable MYFInput with the same structure as each record in the history data list History. The output is an icing susceptibility description output vector ; Step S3, constructing an advance periodic icing feature calculation module MCCFeature, the input of MCCFeature is a periodic list MCCFeatureInput consisting of records of consecutive ZQLength days, and the structure of each record in the periodic list MCCFeatureInput is the same as the record structure of the historical data list History; The module MCCFeature uses MYFVector to process each record in the periodic list MCCFeatureInput separately to obtain the ultra-long lead time periodic icing feature output MCCFeatureOutput; Step S4, constructing an advance periodic icing feature table TZTable using MCCFeature and History, and obtaining an advance periodic identification model SModel based on TZTable; Step S5, input the weather data BTest of the recent ZQLength days in the area to be warned of transmission line icing disaster, each record structure of BTest is the same as the History record structure, and there are a total of ZQLength records; use SModel to warn whether transmission line icing will occur in the area to be warned of transmission line icing disaster in YJLengh days.

[0010] The method of the present invention can establish an indirect correlation between weather and icing occurrence, and provide ultra-long advance warning of icing on transmission lines.

[0011] Specific implementation method 2: This implementation method is different from the specific implementation method 1 in that the specific process of step S1 is as follows: Step S101: input a history data list History of weather and icing occurrence in a region, each record in the history data list History includes weather conditions and icing conditions in one day (i.e., each record is used to record the weather conditions and icing conditions in the history of one day in the region); that is, each record includes the following fields: H00WD: temperature value at 0 o'clock on the day; H00SD: humidity value at 0 o'clock on the day; H06WD: temperature value at 6 o'clock on the day; H06SD: humidity value at 6 o'clock on the day; H12WD: temperature value at 12 o'clock on the day; H12SD: humidity value at 12 o'clock on the day; H18WD: temperature value at 18:00 on the day; H18SD: humidity value at 18:00 on the day; HFS: average wind speed value of the day; HJX: whether there is snowfall on that day, 1 means there is snowfall, 0 means there is no snowfall; HFB: Whether the transmission line is covered with ice on that day, 0 means no ice, 1 means ice; The unit of temperature is Celsius, the unit of humidity is relative humidity, and the unit of wind speed is meter per second. Step S102, input the advance warning days YJLengh, the default value of YJLengh is 3 days; Step S103, input the period to be identified ZQLength, the default value of ZQLength is 5 days; Step S104, JZWD=fetch all records in the history data list History where HFB is equal to 1, and calculate the mean value of the H00WD field in all the fetched records; Step S105, JZSD=fetch all records in the history data list History where HFB is equal to 1, and calculate the mean value of the H00SD field in all the fetched records; Step S106, JZFS=fetch all records in the history data list History where HFB is equal to 1, and calculate the mean value of the HFS field in all the retrieved records; Step S107, JZJX= fetches all records in the history data list History where HFB is equal to 1, and calculates the average value of the HJX field in all the fetched records.

[0012] The other steps and parameters are the same as those in the first embodiment.

[0013] Specific implementation method three: This implementation method is different from specific implementation methods one or two in that the specific process of step S2 is as follows: Step S201, establish an icing susceptibility description acquisition module MYFVector, the input of the module MYFVector is an icing susceptibility description input variable MYFInput with the same structure as each record in the history data list History; Step S202, establish an icing susceptibility description output vector MYFOutput = a vector containing 10 elements, and initialize the values ​​of all elements in the icing susceptibility description output vector MYFOutput to be 0; Step S203, let MYFOutput [1] = (H00WD field of MYFInput - JZWD) / 5; Where MYFOutput [1] is the first element in the icing susceptibility description output vector MYFOutput; Step S204, set MYFOutput [3] = (H06WD field of MYFInput - 3 - JZWD) / 5; Where MYFOutput [3] is the third element in the icing susceptibility description output vector MYFOutput; Step S205, let MYFOutput [5] = (H12WD field of MYFInput - 5 - JZWD) / 5; Where MYFOutput [5] is the fifth element in the icing susceptibility description output vector MYFOutput; Step S206, set MYFOutput [7] = (H18WD field of MYFInput - 3 - JZWD) / 5; Among them, MYFOutput [7] is the 7th element in the output vector MYFOutput describing the icing susceptibility; Step S207, let MYFOutput [2] = (H00SD field of MYFInput - JZSD) / 10; Where MYFOutput [2] is the second element in the icing susceptibility description output vector MYFOutput; Step S208, set MYFOutput [4] = (H06SD field of MYFInput + 10 - JZSD) / 10; Where MYFOutput [4] is the fourth element in the icing susceptibility description output vector MYFOutput; Step S209, set MYFOutput [6] = (H12SD field of MYFInput + 15 - JZSD) / 10; Where MYFOutput [6] is the sixth element in the icing susceptibility description output vector MYFOutput; Step S210, let MYFOutput [8] = (H18SD field of MYFInput + 10 - JZSD) / 10; Where MYFOutput [8] is the 8th element in the output vector MYFOutput describing the icing susceptibility; Step S211, if the HFS field of MYFInput is less than or equal to JZFS, then MYFOutput [9] = 0; If the HFS field of MYFInput is greater than JZFS and the HFS field of MYFInput is less than JZFS+10, then MYFOutput [9] = 0.5, otherwise MYFOutput [9] = 1; Where MYFOutput [9] is the 9th element in the icing susceptibility description output vector MYFOutput; Step S212, let MYFOutput

[10] = HJX field of MYFInput - JZJX; Where MYFOutput

[10] is the 10th element in the icing susceptibility description output vector MYFOutput; Step S213: , where tanh is the hyperbolic tangent value calculated for each element of the vector MYFOutput; Step S214: The calculated value in step S213 As the result output of the icing susceptibility description acquisition module MYFVector.

[0014] The other steps and parameters are the same as those in the first or second embodiment.

[0015] Specific implementation method 4: This implementation method is different from any one of the specific implementation methods 1 to 3 in that the specific process of step S3 is as follows: Step S301, constructing an advance periodic icing feature calculation module MCCFeature, the input of MCCFeature is a periodic list MCCFeatureInput consisting of records of consecutive ZQLength days, and each record structure of the periodic list MCCFeatureInput is the same as the History record structure; Step S302, establish the advance periodic icing feature output MCCFeatureOutput = an empty list; Step S303, advance periodic icing feature single feature counter TZDgCounter=1; Step S304, the first temporary vector of the periodic icing characteristics of the advance amount TZTempVecot1=a vector containing 10 elements, and the values ​​of all elements in the initialization vector are 0; Step S305, the second temporary vector TZTempVecot2 of the periodic icing feature of the advance amount is calculated by using MYFVector, and the input of MYFVector is the TZDgCounterth record in MCCFeatureInput; Step S306, update TZTempVecot1=TZTempVecot2-TZTempVecot1; Step S307, append TZTempVecot2 as a vector to MCCFeatureOutput; Step S308, append TZTempVecot1 as a vector to MCCFeatureOutput; Step S309, set TZTempVecot1=TZTempVecot2; Step S310, set TZDgCounter=TZDgCounter+1; Step S311, if TZDgCounter is less than ZQLength, go to step S305, otherwise go to step S312; Step S312: Output MCCFeatureOutput as the result of MCCFeature.

[0016] The other steps and parameters are the same as those in Specific Embodiments 1 to 3.

[0017] Specific implementation mode 5: This implementation mode is different from the specific implementation modes 1 to 4 in that the specific process of step S4 is as follows: Step S401: construct an advance periodic icing feature table TZTable, which contains two fields: TZInputVector, input sequence characteristics of the advance periodic icing characteristic table; TZDecision, decision characteristics of the advance periodic icing characteristic table; Step S402, ice feature table construction counter TZCounter=1; Step S403, the first temporary variable TZTableTemp1 of the ice feature table temporary variable = calculation is performed using MCCFeature, the input of MCCFeature = all records from the TZCounterth to the TZCounter+ZQLength-1th of History are retrieved; Step S404, the second temporary variable TZTableTemp2 of the ice feature table is stored = taking out the HFB field in the TZCounter+ZQLength-1+YJLenghth record of History; Step S405, create a new row of data for TZTable, the TZInputVector of the row of data is equal to TZTableTemp1, and TZDecision is equal to TZTableTemp2; Step S406, set TZCounter=TZCounter+1; Step S407, if TZCounter+ZQLength-1+YJLengh is less than the number of data entries in History, go to step S403, otherwise go to step S408; Step S408, SModel=establishes a long short-term memory network (i.e., LSTM). The input of SModel is a sequence of ZQLength vectors, and each vector of the sequence is 10-dimensional. The output of SModel is a decision output. Output 1 indicates that the prediction result is that an ice disaster will occur after YJLengh days in the future, and output 0 indicates that the prediction result is that an ice disaster will not occur after YJLengh days in the future. Step S409, correspond the input of SModel to the TZInputVector field of TZTable, and correspond the output of SModel to the TZDecision field of TZTable; Step S410: Use the data in TZTable to train SModel, and the trained SModel is used to predict icing disasters.

[0018] The other steps and parameters are the same as those in Specific Embodiments 1 to 4.

[0019] The SModel in this implementation can be trained using weather and ice condition data from multiple regions, but the input in each set of training data must be input for consecutive ZQLength days in the same region.

[0020] Specific implementation method 6: This implementation method is different from the specific implementation methods 1 to 5 in that the specific process of step S5 is as follows: Step S501, inputting the weather data BTest of the recent ZQLength days in the area to be warned of the icing disaster of the transmission line; Step S502, the periodic ice feature to be decided DJCFeature=calculated using MCCFeature, the input of MCCFeature=BTest; Step S503, input DJCFeature into SModel, obtain the decision output of SModel and store the decision output into the decision variable DecParam; Step S504: If DecParam is equal to 1, go to step S505; otherwise, go to step S506; Step S505: A special period that typically causes icing disasters is found, and a transmission line icing disaster will occur in the area to be warned in YJLengh days, and the process goes to step S507; Step S506: No special cycle that typically causes icing disasters is found, and the transmission line icing disaster warning area will not have icing disasters after YJLengh days, and the process goes to step S507; Step S507: The whole process ends.

[0021] The other steps and parameters are the same as those in Specific Implementation Methods 1 to 5.

[0022] Example In order to test the effectiveness of the method of the present invention, the weather and icing conditions in 2022 and 2023 in a mountainous area in the north where icing of power transmission lines is prone to occur were introduced; the method of the present invention and some traditional methods were used to predict the icing conditions 3 days in advance, and the results were compared as shown in Table 1: Table 1 method There were 17 ice cover events in the past two years, and the number of successful predictions was Prediction accuracy (%) Number of false positives Method of the present invention 16 94.12 1 Method for detecting data exceeding threshold 1 5.88 0 Micro-weather + LSTM deep neural network 8 47.06 27 Micro-weather + Transformer deep neural network 10 58.82 47 It can be seen from the results that the method of the present invention can provide ultra-long advance warning for icing, while other methods try their best to find the direct correlation between weather and icing, and have low accuracy and high number of false alarms in ultra-long advance warning.

[0023] The above calculation examples of the present invention are only used to explain the calculation model and calculation process of the present invention in detail, and are not intended to limit the implementation methods of the present invention. For ordinary technicians in the relevant field, other different forms of changes or modifications can be made based on the above description. It is impossible to list all the implementation methods here. All obvious changes or modifications derived from the technical solution of the present invention are still within the scope of protection of the present invention.

Claims

1. A transmission line icing warning method based on weather factors, characterized in that: The method specifically comprises the following steps: Step S1, inputting a list of historical data of weather and icing occurrence in a region History, the number of early warning days YJLengh and the length of the period to be identified ZQLength; According to the historical data list History, the mean temperature JZWD, mean humidity JZSD, mean wind speed JZFS and snowfall correlation JZJX in the statistical area under ice cover conditions; Step S2: Establish an icing susceptibility description acquisition module MYFVector. The input of the module MYFVector is an icing susceptibility description input variable MYFInput with the same structure as each record in the history data list History. The output is an icing susceptibility description output vector ; Step S3, constructing an advance periodic icing feature calculation module MCCFeature, the input of MCCFeature is a periodic list MCCFeatureInput consisting of records of consecutive ZQLength days, and the structure of each record in the periodic list MCCFeatureInput is the same as the record structure of the historical data list History; The module MCCFeature uses MYFVector to process each record in the periodic list MCCFeatureInput separately to obtain the advance periodic icing feature output MCCFeatureOutput; Step S4, constructing an advance periodic icing feature table TZTable using MCCFeature and History, and obtaining an advance periodic identification model SModel based on TZTable; Step S5, input the weather data BTest of the recent ZQLength days in the area to be warned of transmission line icing disaster, and use SModel to warn whether transmission line icing will occur in the area to be warned of transmission line icing disaster in YJLengh days.

2. The method for early warning of ice coating on transmission lines based on weather factors according to claim 1, characterized in that: The specific process of step S1 is as follows: Step S101: input a history data list History of weather and ice occurrence in a region. Each record in the history data list History includes weather conditions and ice conditions in one day; that is, each record includes the following fields: H00WD: temperature value at 0 o'clock on the day; H00SD: humidity value at 0 o'clock on the day; H06WD: temperature value at 6 o'clock on the day; H06SD: humidity value at 6 o'clock on the day; H12WD: temperature value at 12 o'clock on the day; H12SD: humidity value at 12 o'clock on the day; H18WD: temperature value at 18:00 on the day; H18SD: humidity value at 18:00 on the day; HFS: average wind speed value of the day; HJX: whether there is snowfall on that day, 1 means there is snowfall, 0 means there is no snowfall; HFB: Whether the transmission line is covered with ice on that day, 0 means no ice, 1 means ice; The unit of temperature is Celsius, the unit of humidity is relative humidity, and the unit of wind speed is meter per second. Step S102, input the number of advance warning days YJLengh; Step S103, input the period ZQLength to be identified; Step S104, JZWD=fetch all records in the history data list History where HFB is equal to 1, and calculate the mean value of the H00WD field in all the fetched records; Step S105, JZSD=fetch all records in the history data list History where HFB is equal to 1, and calculate the mean value of the H00SD field in all the fetched records; Step S106, JZFS=fetch all records in the history data list History where HFB is equal to 1, and calculate the mean value of the HFS field in all the retrieved records; Step S107, JZJX= fetches all records in the history data list History where HFB is equal to 1, and calculates the average value of the HJX field in all the fetched records.

3. The method for early warning of ice coating on transmission lines based on weather factors according to claim 2 is characterized in that: The specific process of step S2 is: Step S201, establish an icing susceptibility description acquisition module MYFVector, the input of the module MYFVector is an icing susceptibility description input variable MYFInput with the same structure as each record in the history data list History; Step S202, establish an icing susceptibility description output vector MYFOutput = a vector containing 10 elements, and initialize the values ​​of all elements in the icing susceptibility description output vector MYFOutput to be 0; Step S203, let MYFOutput [1] = (H00WD field of MYFInput - JZWD) / 5; Where MYFOutput [1] is the first element in the icing susceptibility description output vector MYFOutput; Step S204, set MYFOutput [3] = (H06WD field of MYFInput - 3 - JZWD) / 5; Where MYFOutput [3] is the third element in the icing susceptibility description output vector MYFOutput; Step S205, let MYFOutput [5] = (H12WD field of MYFInput - 5 - JZWD) / 5; Where MYFOutput [5] is the fifth element in the icing susceptibility description output vector MYFOutput; Step S206, set MYFOutput [7] = (H18WD field of MYFInput - 3 - JZWD) / 5; Among them, MYFOutput [7] is the 7th element in the output vector MYFOutput describing the icing susceptibility; Step S207, let MYFOutput [2] = (H00SD field of MYFInput - JZSD) / 10; Where MYFOutput [2] is the second element in the icing susceptibility description output vector MYFOutput; Step S208, set MYFOutput [4] = (H06SD field of MYFInput + 10 - JZSD) / 10; Where MYFOutput [4] is the fourth element in the icing susceptibility description output vector MYFOutput; Step S209, set MYFOutput [6] = (H12SD field of MYFInput + 15 - JZSD) / 10; Where MYFOutput [6] is the sixth element in the icing susceptibility description output vector MYFOutput; Step S210, let MYFOutput [8] = (H18SD field of MYFInput + 10 - JZSD) / 10; Where MYFOutput [8] is the 8th element in the output vector MYFOutput describing the icing susceptibility; Step S211, if the HFS field of MYFInput is less than or equal to JZFS, then MYFOutput [9] = 0; If the HFS field of MYFInput is greater than JZFS and the HFS field of MYFInput is less than JZFS+10, then MYFOutput[9]=0.5, otherwise MYFOutput[9]=1; Where MYFOutput [9] is the 9th element in the icing susceptibility description output vector MYFOutput; Step S212, let MYFOutput [10] = HJX field of MYFInput - JZJX; Where MYFOutput [10] is the 10th element in the icing susceptibility description output vector MYFOutput; Step S213: , where tanh is the hyperbolic tangent value calculated for each element of the vector MYFOutput; Step S214: The calculated value in step S213 As the result output of the icing susceptibility description acquisition module MYFVector.

4. The method for early warning of ice coating on transmission lines based on weather factors according to claim 3 is characterized in that: The specific process of step S3 is: Step S301, constructing an advance periodic icing feature calculation module MCCFeature, the input of MCCFeature is a periodic list MCCFeatureInput consisting of records of consecutive ZQLength days, and each record structure of the periodic list MCCFeatureInput is the same as the History record structure; Step S302, establish the advance periodic icing feature output MCCFeatureOutput = an empty list; Step S303, advance periodic icing feature single feature counter TZDgCounter=1; Step S304, the first temporary vector of the periodic icing characteristics of the advance amount TZTempVecot1=a vector containing 10 elements, and the values ​​of all elements in the initialization vector are 0; Step S305, the second temporary vector TZTempVecot2 of the periodic icing feature of the advance amount is calculated by using MYFVector, and the input of MYFVector is the TZDgCounterth record in MCCFeatureInput; Step S306, update TZTempVecot1=TZTempVecot2-TZTempVecot1; Step S307, append TZTempVecot2 as a vector to MCCFeatureOutput; Step S308, append TZTempVecot1 as a vector to MCCFeatureOutput; Step S309, set TZTempVecot1=TZTempVecot2; Step S310, set TZDgCounter=TZDgCounter+1; Step S311, if TZDgCounter is less than ZQLength, go to step S305, otherwise go to step S312; Step S312: Output MCCFeatureOutput as the result of MCCFeature.

5. The method for early warning of ice coating on transmission lines based on weather factors according to claim 4 is characterized in that: The specific process of step S4 is as follows: Step S401: construct an advance periodic icing feature table TZTable, which contains two fields: TZInputVector, input sequence characteristics of the advance periodic icing characteristic table; TZDecision, decision characteristics of the advance periodic icing characteristic table; Step S402, ice feature table construction counter TZCounter=1; Step S403, the first temporary variable TZTableTemp1 of the ice feature table temporary variable = calculation is performed using MCCFeature, the input of MCCFeature = all records from the TZCounterth to the TZCounter+ZQLength-1th of History are retrieved; Step S404, the second temporary variable TZTableTemp2 of the ice feature table is stored = taking out the HFB field in the TZCounter+ZQLength-1+YJLenghth record of History; Step S405, create a new row of data for TZTable, the TZInputVector of the row of data is equal to TZTableTemp1, and TZDecision is equal to TZTableTemp2; Step S406, set TZCounter=TZCounter+1; Step S407, if TZCounter+ZQLength-1+YJLengh is less than the number of data entries in History, go to step S403, otherwise go to step S408; Step S408, SModel=establishes a long short-term memory network, the input of SModel is a sequence of ZQLength vectors, each vector of the sequence is 10-dimensional; the output of SModel is a decision output, output 1 indicates that the prediction result is that an ice disaster will occur after YJLengh days in the future, and output 0 indicates that the prediction result is that an ice disaster will not occur after YJLengh days in the future; Step S409, correspond the input of SModel to the TZInputVector field of TZTable, and correspond the output of SModel to the TZDecision field of TZTable; Step S410: Use the data in TZTable to train SModel, and the trained SModel is used to predict icing disasters.

6. The method for early warning of ice coating on transmission lines based on weather factors according to claim 5, characterized in that: The specific process of step S5 is as follows: Step S501, inputting the weather data BTest of the recent ZQLength days in the area to be warned of the icing disaster of the transmission line; Step S502, the periodic ice feature to be decided DJCFeature=calculated using MCCFeature, the input of MCCFeature=BTest; Step S503, input DJCFeature into SModel, obtain the decision output of SModel and store the decision output into the decision variable DecParam; Step S504: If DecParam is equal to 1, go to step S505; otherwise, go to step S506; Step S505: In the area to be warned of the icing disaster of the transmission line, the icing disaster of the transmission line will occur in YJLengh days, and the process goes to step S507; Step S506: If the area to be warned of the icing disaster of the transmission line will not experience the icing disaster of the transmission line after YJLengh days, then go to step S507; Step S507: The whole process ends.

Images