Icing thickness prediction method and device for overhead line, equipment and storage medium

Abstract

The invention relates to an icing thickness prediction method and device for an overhead line, equipment and a storage medium. The method comprises the following steps: acquiring time sequence monitoring data at a monitoring point on an overhead line in a target time period; wherein the time sequence monitoring data comprises meteorological monitoring data and line state monitoring data at each sampling moment in the target time period; time windows with different lengths are adopted, and multi-scale time features of the time sequence monitoring data are extracted; selecting a current icing state of the overhead line from the candidate icing states according to the time sequence monitoring data; wherein each candidate icing state corresponds to one thickness prediction model; and based on the thickness prediction model corresponding to the current icing state, according to the multi-scale time characteristics, determining the predicted icing thickness of the overhead line in the future time period. By adopting the method, the icing thickness of the overhead line can be accurately predicted.

Description

Technical Field

[0001] The present application relates to the technical field of distribution network lines, and in particular to a method, device, equipment and storage medium for predicting ice thickness of overhead lines. Background Art

[0002] Icing on overhead lines poses a serious threat to power systems, particularly in cold climates. It can lead to increased line loads, reduced mechanical strength, and severe consequences such as line breakage and collapse. Accurately predicting the thickness of ice on overhead lines is crucial for safe grid operation and disaster prevention and mitigation. With the development of the Internet of Things (IoT), power systems have deployed a large number of meteorological monitoring devices, generating massive amounts of real-time data daily. This high-frequency, multi-dimensional data provides the foundation for icing prediction, but it also presents data processing challenges.

[0003] Existing methods for predicting icing on overhead lines often use a static batch processing model, which fails to capture the dynamic evolution of icing formation. Furthermore, existing methods typically treat meteorological data as independent samples, ignoring the variations in meteorological parameters over different timescales. This results in low prediction accuracy during the initial stages of icing formation and during transitions. Summary of the Invention

[0004] Based on this, it is necessary to provide a method, device, equipment and storage medium for predicting the ice thickness of overhead lines to address the above technical problems, which can accurately predict the ice thickness of overhead lines.

[0005] In a first aspect, the present application provides a method for predicting ice thickness of overhead lines, comprising:

[0006] Acquire time series monitoring data at monitoring points on the overhead line within a target period; wherein the time series monitoring data includes meteorological monitoring data and line status monitoring data at each sampling time within the target period;

[0007] Using time windows of different lengths to extract multi-scale time features of the time series monitoring data;

[0008] selecting a current icing state of the overhead line from candidate icing states according to the time series monitoring data; wherein each candidate icing state corresponds to a thickness prediction model;

[0009] Based on the thickness prediction model corresponding to the current icing state and according to the multi-scale time characteristics, the predicted icing thickness of the overhead line in a future time period is determined.

[0010] In one embodiment, extracting multi-scale time features of the time series monitoring data using time windows of different lengths includes:

[0011] Using time windows of different lengths to extract the time series monitoring data to obtain data sequences of different lengths;

[0012] For each data sequence of length, statistical analysis is performed on the meteorological monitoring data and line status monitoring data at the sampling time contained in the data sequence to obtain statistical parameters corresponding to the data sequence;

[0013] The multi-scale time characteristics of the time series monitoring data are determined according to the statistical parameters corresponding to each data sequence.

[0014] In one embodiment, the line status monitoring data includes at least the original ice thickness at the endpoint of the overhead line;

[0015] The step of selecting the current icing state of the overhead line from candidate icing states according to the time series monitoring data includes:

[0016] Determining a weighted difference of the original ice thickness at the endpoints according to the original ice thickness at each sampling moment within the target period;

[0017] The current icing state of the overhead line is selected from candidate icing states according to the weighted difference and the meteorological monitoring data at the end sampling moment in the target period.

[0018] In one embodiment, the candidate icing states include an ice-free state, an initial icing state, a stable growth state, and a melting state; the meteorological monitoring data includes at least the temperature and humidity of the environment where the overhead line is located.

[0019] The selecting the current icing state of the overhead line from candidate icing states based on the weighted difference and the meteorological monitoring data at the end sampling time within the target period includes:

[0020] determining whether the overhead line meets ice growth conditions based on the temperature and humidity at the end sampling time within the target period;

[0021] If the original ice thickness at the endpoint at the last sampling moment is less than the initial ice thickness threshold, the weighted difference is greater than zero, and the overhead line meets the ice growth condition, then determining that the current ice state of the overhead line is the initial ice state;

[0022] If the original ice thickness at the endpoint at the last sampling moment is greater than or equal to the initial ice thickness threshold, the weighted difference is greater than zero, and the overhead line meets the ice growth condition, then determining that the current ice state of the overhead line is a stable growth state;

[0023] If the weighted difference is less than zero and the overhead line does not meet the ice growth condition, determining that the current ice state of the overhead line is a melting state;

[0024] If the original ice thickness of the endpoint is less than the minimum detectable ice thickness and the overhead line does not meet the ice growth condition, it is determined that the current ice state of the overhead line is an ice-free state.

[0025] In one embodiment, the determining, based on the thickness prediction model corresponding to the current icing state and according to the multi-scale time characteristics, the predicted icing thickness of the overhead line in a future period includes:

[0026] Inputting the multi-scale time characteristics into a thickness prediction model corresponding to the current ice cover state to obtain an initial predicted thickness;

[0027] The initial predicted thickness is adjusted according to an ice adjustment parameter that characterizes an ice growth law of the current ice state to obtain a predicted ice thickness of the overhead line in a future period.

[0028] In one embodiment, the method further comprises:

[0029] Obtaining actual ice thickness of the overhead line in a future period;

[0030] Determining a prediction error of a thickness prediction model corresponding to the current ice cover state according to the actual ice cover thickness and the predicted ice cover thickness;

[0031] According to the prediction error, the thickness prediction model corresponding to the current ice cover state is adjusted.

[0032] In one embodiment, different thickness prediction models use different initialization strategies and activation functions during the training process, and the initialization strategy and activation function used by any thickness prediction model during the training process are determined according to the candidate icing state corresponding to the thickness prediction model.

[0033] In a second aspect, the present application further provides an apparatus for predicting ice thickness of an overhead line, comprising:

[0034] A data acquisition module, configured to acquire time-series monitoring data at monitoring points on overhead lines within a target period; wherein the time-series monitoring data includes meteorological monitoring data and line status monitoring data at each sampling moment within the target period;

[0035] A feature extraction module, configured to extract multi-scale time features of the time series monitoring data using time windows of different lengths;

[0036] a state selection module, configured to select a current icing state of the overhead line from candidate icing states based on the time series monitoring data; wherein each candidate icing state corresponds to a thickness prediction model;

[0037] A thickness prediction module is used to determine the predicted ice thickness of the overhead line in a future period based on the thickness prediction model corresponding to the current ice state and according to the multi-scale time characteristics.

[0038] In a third aspect, the present application further provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0039] Acquire time series monitoring data at monitoring points on the overhead line within a target period; wherein the time series monitoring data includes meteorological monitoring data and line status monitoring data at each sampling time within the target period;

[0040] Using time windows of different lengths to extract multi-scale time features of the time series monitoring data;

[0041] selecting a current icing state of the overhead line from candidate icing states according to the time series monitoring data; wherein each candidate icing state corresponds to a thickness prediction model;

[0042] Based on the thickness prediction model corresponding to the current icing state and according to the multi-scale time characteristics, the predicted icing thickness of the overhead line in a future time period is determined.

[0043] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:

[0044] Acquire time series monitoring data at monitoring points on the overhead line within a target period; wherein the time series monitoring data includes meteorological monitoring data and line status monitoring data at each sampling time within the target period;

[0045] Using time windows of different lengths to extract multi-scale time features of the time series monitoring data;

[0046] selecting a current icing state of the overhead line from candidate icing states according to the time series monitoring data; wherein each candidate icing state corresponds to a thickness prediction model;

[0047] Based on the thickness prediction model corresponding to the current icing state and according to the multi-scale time characteristics, the predicted icing thickness of the overhead line in a future time period is determined.

[0048] In a fifth aspect, the present application further provides a computer program product, comprising a computer program, which, when executed by a processor, implements the following steps:

[0049] Acquire time series monitoring data at monitoring points on the overhead line within a target period; wherein the time series monitoring data includes meteorological monitoring data and line status monitoring data at each sampling time within the target period;

[0050] Using time windows of different lengths to extract multi-scale time features of the time series monitoring data;

[0051] selecting a current icing state of the overhead line from candidate icing states according to the time series monitoring data; wherein each candidate icing state corresponds to a thickness prediction model;

[0052] Based on the thickness prediction model corresponding to the current icing state and according to the multi-scale time characteristics, the predicted icing thickness of the overhead line in a future time period is determined.

[0053] The above-mentioned overhead line ice thickness prediction method, device, equipment and storage medium, after obtaining the time series monitoring data of meteorological monitoring data and line status monitoring data at the monitoring points on the overhead line within the target period, including the meteorological monitoring data and line status monitoring data at each sampling time within the target period, use time windows of different lengths to extract the multi-scale time characteristics of the time series monitoring data, comprehensively consider the characteristics of different scales and multiple dimensions, effectively capture the law of change of meteorological parameters over time, and provide key data support for the dynamic prediction of the icing process; further, according to the time series monitoring data, the current icing state of the overhead line is selected from the candidate icing states, and each candidate icing state corresponds to a thickness prediction model. By formulating targeted thickness prediction models for different icing states, accurate prediction of different icing stages is achieved; finally, based on the thickness prediction model corresponding to the current icing state, according to the multi-scale time characteristics, the predicted icing thickness of the overhead line in the future period is determined, thereby ensuring the accuracy of the determined predicted icing thickness. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following briefly introduces the drawings required for use in the embodiments or related technical descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0055] Figure 1 A diagram illustrating an application environment of a method for predicting ice thickness of overhead lines in one embodiment;

[0056] Figure 21 is a flow chart of a method for predicting ice thickness of overhead lines in one embodiment;

[0057] Figure 3 A schematic diagram of a process for determining multi-scale temporal features in one embodiment;

[0058] Figure 4 Schematic diagram of a process for determining the current ice cover state in one embodiment;

[0059] Figure 5 A schematic diagram of a process for adjusting an output result of a thickness prediction model in one embodiment;

[0060] Figure 6 Schematic diagram of a process for adjusting a thickness prediction model in one embodiment;

[0061] Figure 7 1 is a flow chart of a method for predicting ice thickness of overhead lines in another embodiment;

[0062] Figure 8 is a structural block diagram of an apparatus for predicting ice thickness of overhead lines in one embodiment;

[0063] Figure 9 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0064] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0065] The method for predicting ice thickness of overhead lines provided in the embodiment of the present application can be applied to Figure 1In the application environment shown. The prediction device 101 is used to predict the ice thickness of the overhead line. The prediction device 101 can be a server or a terminal with powerful computing capabilities. The monitoring device 102 is installed at the monitoring point of the overhead line and is used to collect meteorological monitoring data and line status monitoring data. The monitoring device 102 can be a sensor. Optionally, the prediction device 101 obtains the time series monitoring data collected by the monitoring device 102 installed at the monitoring point on the overhead line during the target period. The time series monitoring data includes meteorological monitoring data and line status monitoring data at each sampling moment during the target period. Time windows of different lengths are used to extract multi-scale time features of the time series monitoring data. Based on the time series monitoring data, the current ice state of the overhead line is selected from the candidate ice states. Each candidate ice state corresponds to a thickness prediction model. Based on the thickness prediction model corresponding to the current ice state, the predicted ice thickness of the overhead line in the future period is determined according to the multi-scale time features.

[0066] In an exemplary embodiment, Figure 2 As shown in the figure, a method for predicting ice thickness of overhead lines is provided. Figure 1 The prediction device 101 in FIG. 1 is used as an example to illustrate the method, which specifically includes the following steps:

[0067] S201, acquiring time-series monitoring data at monitoring points on overhead lines within a target period.

[0068] Among them, the target time period is any preset time period, which can be a time period within any current time length; the overhead line can be an overhead line in any area of ​​the power system; the monitoring point is a preset monitoring point on the overhead line, and each monitoring point is installed with a monitoring device for monitoring the meteorological data at the monitoring point of the overhead line and the line status data of the overhead line. It can be understood that the monitoring device at each monitoring point can be one or an array of multiple monitoring devices; the time series monitoring data is the time series data collected by the monitoring device at the monitoring point. In the embodiment of the present application, the time series monitoring device includes but is not limited to the meteorological monitoring data and line status monitoring data at each sampling time within the target time period.

[0069] Meteorological monitoring data refers to the meteorological data of the environment where the monitoring point of the overhead line is located, and the meteorological monitoring data includes but is not limited to temperature, humidity and wind speed; line status monitoring data refers to the status of the overhead line at the monitoring point, and the line status monitoring data includes but is not limited to the height of the overhead line and the original thickness of ice cover at the endpoint, where the original thickness of ice cover at the endpoint is the original measured value of ice cover at the endpoint.

[0070] Optionally, the meteorological data and line status data at the monitoring points on the overhead lines collected by the monitoring equipment during the target period can be obtained from the monitoring equipment installed at the monitoring points of the overhead lines, and the meteorological data can be used as meteorological monitoring data and the line status data can be used as line status monitoring data; further, the meteorological monitoring data and the line status monitoring data can be used together as initial monitoring data.

[0071] For example, at this monitoring point, the monitoring data acquired at any sampling time within the target period can be expressed as: .in, For monitoring points exist Monitoring data at all times; For monitoring points exist The temperature of the moment; For monitoring points exist Humidity at all times; For monitoring points exist Wind speed at the moment; For monitoring points exist The height of the overhead lines at all times; For monitoring points exist The original thickness of ice cover at the end point of the overhead line at that moment.

[0072] Furthermore, the time series monitoring data at the monitoring points on the overhead line during the target period can be expressed as: .in, It is time series monitoring data; The maximum time length of the target period.

[0073] It should be noted that since the directly acquired time series monitoring data may contain missing values ​​or abnormal values, after obtaining the initial monitoring data, the initial monitoring data can be preprocessed to obtain the time series monitoring data. For example, the missing values ​​in the initial monitoring data can first be processed using a time interpolation method, which can be specifically expressed as the following formula (1):

[0074] (1)

[0075] in, ; and Is the missing value moment The most recent valid data moment.

[0076] Furthermore, the Z-score method based on sliding window is used to detect outliers:

[0077] (2)

[0078] in, and Characteristics In the time window The mean and standard deviation within , is the window size, set to 24 (24 hours).

[0079] when When , the value is considered as an abnormal value and replaced using the following method:

[0080] (3)

[0081] Next, the cleaned data is standardized:

[0082] (4)

[0083] in, and Characteristics The mean and standard deviation of .

[0084] S202 , using time windows of different lengths to extract multi-scale time features of the time series monitoring data.

[0085] Among them, time windows of different lengths can be classified into windows of different time length levels, which are not limited by the task here. For example, time windows can be classified into short time length windows, medium time length windows and long time length windows; further, the time length of short time length windows can be 1 hour, 3 hours and 6 hours, the time length of medium time length windows can be 12 hours and 24 hours, and the time length of long time length windows can be 48 hours, 72 hours and 168 hours. Multi-scale time features are features that characterize multi-dimensional time in time series monitoring data, that is, features of different time dimensions.

[0086] Optionally, pre-set time windows of different lengths may be used to extract time features of different time windows from the time series monitoring data as multi-scale time features.

[0087] S203: Select the current icing state of the overhead line from the candidate icing states according to the time series monitoring data.

[0088] Each candidate icing state corresponds to a thickness prediction model. In an embodiment of the present application, candidate icing states include, but are not limited to, an ice-free state, an initial icing state, a stable growth state, and a melting state. To ensure that the thickness prediction model corresponding to each candidate icing state is more targeted, different thickness prediction models use different initialization strategies and activation functions during training. The initialization strategy and activation function used by any thickness prediction model during training are determined based on the candidate icing state corresponding to the thickness prediction model. In addition, different thickness prediction models can all be constructed using the same neural network model, such as an extreme learning machine model.

[0089] For each candidate icing state ,in, Indicates no ice state. Indicates the initial icing state, Indicates a stable growth state. Indicates the melting state and constructs the corresponding thickness prediction model:

[0090] (5)

[0091] in, is the status The input feature matrix contains vector. It is a monitoring point Previous Status The number of samples, is the feature dimension, which is composed of the merged feature vector The length is determined; is the hidden layer output matrix, is the status The number of hidden layer nodes; is the target ice thickness; is the input weight matrix; is the bias vector; is the output weight vector.

[0092] The thickness prediction model corresponding to each candidate ice cover state uses a different initialization strategy. Specifically, the number of hidden layer nodes is set as follows:

[0093] (6)

[0094] Input weight initialization:

[0095] (7)

[0096] in, is the status The weight initialization range.

[0097] Bias vector initialization:

[0098] (8)

[0099] in is the status The bias initialization range.

[0100] Select the optimal activation function for the thickness prediction model corresponding to different candidate icing states:

[0101] (9)

[0102] Applied to hidden layer output calculation:

[0103] (10)

[0104] in, Is the length of A vector of all 1s, Represents the vector outer product.

[0105] It should be noted that different thickness prediction models can be trained and calculated using the Storm stream processing framework, and implement incremental learning capabilities, enabling the system to learn from continuously arriving new data and continuously optimize model parameters without retraining.

[0106] Optionally, preset rules of monitoring data of overhead lines under different candidate icing states can be analyzed based on historical data. Further, the rules of time series monitoring data can be analyzed, and a comparative analysis can be performed based on the rules obtained from the analysis and the preset rules to select the current icing state of the overhead line from the candidate icing states.

[0107] S204: Based on a thickness prediction model corresponding to the current icing state and according to multi-scale time characteristics, a predicted icing thickness of the overhead line in a future period is determined.

[0108] The predicted ice thickness is the predicted ice thickness at the overhead line monitoring point.

[0109] Optionally, the multi-scale time features may be input into a thickness prediction model corresponding to the current icing state, so that the thickness prediction model analyzes the multi-scale time features and predicts the ice thickness of the overhead line in the future period.

[0110] It should be noted that each thickness prediction model can be trained and updated in real time. After the training is completed in the previous period, after obtaining the data collected in the next period, the model parameters of the thickness prediction model will be updated according to the data collected in the next period. Status , when new data batch Upon arrival (of which, is the corresponding ice thickness), a new input matrix and target vector are formed:

[0111] (11)

[0112] Calculate the new hidden layer output matrix:

[0113] (12)

[0114] Update the covariance matrix:

[0115] (13)

[0116] Update output weights:

[0117] (14)

[0118] In the above-mentioned method for predicting the ice thickness of overhead lines, after obtaining the time series monitoring data of meteorological monitoring data and line status monitoring data at the monitoring points on the overhead lines within the target period, including the meteorological monitoring data at each sampling time within the target period, time windows of different lengths are used to extract the multi-scale time characteristics of the time series monitoring data, comprehensively considering the characteristics of different scales and multiple dimensions, effectively capturing the law of change of meteorological parameters over time, and providing key data support for the dynamic prediction of the icing process; further, based on the time series monitoring data, the current icing state of the overhead line is selected from the candidate icing states, and each candidate icing state corresponds to a thickness prediction model. By formulating targeted thickness prediction models for different icing states, accurate predictions for different icing stages are achieved; finally, based on the thickness prediction model corresponding to the current icing state, according to the multi-scale time characteristics, the predicted icing thickness of the overhead line in the future period is determined, thereby ensuring the accuracy of the determined predicted icing thickness.

[0119] Optionally, in order to ensure the accuracy of the acquired multi-scale time features, in an exemplary embodiment, as Figure 3 As shown, a method for determining multi-scale time features is provided to refine the above S202, specifically comprising the following steps:

[0120] S301 , extracting time series monitoring data using time windows of different lengths to obtain data sequences of different lengths.

[0121] Each data sequence includes meteorological monitoring data and line status monitoring data at each sampling moment within the time length corresponding to the data sequence.

[0122] Optionally, time windows of different lengths may be used to extract time series monitoring data by sliding the time windows to obtain data sequences of different lengths.

[0123] S302 : For each data sequence of length, statistical analysis is performed on the meteorological monitoring data and line status monitoring data at the sampling moments contained in the data sequence to obtain statistical parameters corresponding to the data sequence.

[0124] Statistical parameters are parameters that characterize the mathematical characteristics of a data sequence. In the embodiment of this application, statistical parameters include, but are not limited to, mean, standard deviation, maximum value, minimum value, rate of change, and fluctuation frequency. It should be noted that the statistical parameters corresponding to each data sequence include the statistical parameters of each feature in the data sequence, such as temperature, humidity, wind speed, overhead line height, and original ice thickness at endpoints.

[0125] Optionally, for a data sequence of each length, a statistical analysis is performed on the meteorological monitoring data and line status monitoring data at the sampling moments contained in the data sequence, and statistical parameters of each feature in the data sequence are calculated respectively.

[0126] For example, the mean can be calculated by the following formula (15):

[0127] (15)

[0128] in, Indicates the length of the time window.

[0129] The standard deviation can be calculated using the following formula (16):

[0130] (16)

[0131] The maximum value can be calculated by the following formula (17):

[0132] (17)

[0133] The minimum value can be calculated by the following formula (18):

[0134] (18)

[0135] The rate of change can be calculated using the following formula (19):

[0136] (19)

[0137] The main frequency can be calculated by Fourier transform, which is specifically expressed as:

[0138] (20)

[0139] S303: Determine the multi-scale time characteristics of the time series monitoring data according to the statistical parameters corresponding to each data sequence.

[0140] Optionally, the statistical parameters of data sequences of different lengths can be combined into a unified feature vector, for example:

[0141] (twenty one)

[0142] (twenty two)

[0143] (twenty three)

[0144] in, A sequence of statistical parameters representing a data series of short time length; Represents the mean of a data series with a time length of 1 in a short time length.

[0145] Furthermore, the final multi-scale time feature of the time series monitoring data is a combination of all the above feature vectors.

[0146] In this embodiment, by performing statistical analysis on the meteorological monitoring data and line status monitoring data at the sampling moments contained in the data sequences of each length, the mathematical characteristics of the meteorological monitoring data and line status monitoring data in each data sequence can be fully analyzed, thereby ensuring the accuracy of the constructed multi-scale time features.

[0147] Optionally, the line status monitoring data includes at least the original ice thickness of the endpoint of the overhead line; in this case, in order to ensure the accuracy of the current ice status determined, in one embodiment, as shown in FIG. Figure 4 As shown, a method for determining the current ice cover state is provided to refine the above S203, specifically comprising the following steps:

[0148] S401: Determine a weighted difference of the original ice thickness at the endpoints according to the original ice thickness at each sampling moment within a target period.

[0149] The weighted difference represents the difference between the original ice thickness at the end point of the target period at the last sampling moment and the original ice thickness at the sampling moment before the last sampling moment.

[0150] Optionally, the process of determining the weighted difference of the original ice thickness at the endpoints at each sampling time within the target period can be specifically expressed as follows:

[0151] (twenty four)

[0152] in, is a decreasing weight, the superscript Indicates weighting.

[0153] S402 : Selecting a current icing state of the overhead line from candidate icing states based on the weighted difference and the meteorological monitoring data at the end sampling time of the target period.

[0154] Optionally, a judgment rule may be set in advance for each candidate icing state, and based on the judgment rule, the current icing state of the overhead line may be selected from the candidate icing states according to the weighted difference and the meteorological monitoring data at the end sampling moment.

[0155] For example, a method for determining whether an overhead line meets ice growth conditions can be introduced, and the current ice state of the overhead line can be selected from candidate ice states by combining weighted difference and meteorological monitoring data at the last sampling time. Specifically, whether the overhead line meets ice growth conditions is determined based on the temperature and humidity at the last sampling time within the target period. Specifically, if the temperature and humidity at the last sampling time within the target period meet the following conditions, it can be determined that the overhead line meets ice growth conditions:

[0156] (25)

[0157] in, is the frost point temperature, for example, it can be set to 0; For critical humidity, for example, it can be set to 85%.

[0158] Furthermore, if the original ice thickness at the endpoint at the final sampling moment is less than the initial ice thickness threshold, the weighted difference is greater than zero, and the overhead line meets the ice growth condition, then the current ice state of the overhead line is determined to be the initial ice state. This is shown in the following formula (26):

[0159] (26)

[0160] in, Initial ice thickness threshold; is the value in the above formula (24) The value calculated when the value is 6.

[0161] Optionally, if the original ice thickness at the endpoint at the final sampling moment is greater than or equal to the initial ice thickness threshold, the weighted difference is greater than zero, and the overhead line meets the ice growth condition, then the current ice state of the overhead line is determined to be a stable growth state. A weighted difference greater than zero indicates that the original ice thickness at the endpoint is increasing. This is shown in the following formula (27):

[0162] (27)

[0163] Optionally, if the weighted difference is less than zero and the overhead line does not meet the ice growth condition, the current ice state of the overhead line is determined to be a melting state. A weighted difference less than zero indicates that the original ice thickness at the endpoint is decreasing. This is shown in the following formula (28):

[0164] (28)

[0165] Optionally, if the original ice thickness at the endpoint is less than the minimum detectable ice thickness and the overhead line does not meet the ice growth conditions, the current ice state of the overhead line is determined to be ice-free. Specifically, as shown in the following formula (29):

[0166] (29)

[0167] in, is the minimum detectable ice thickness.

[0168] It is understandable that by introducing the judgment rule of ice growth conditions and formulating a targeted judgment rule for each candidate ice state, the current ice state can be determined quickly and accurately.

[0169] In this embodiment, by introducing the weighted difference of the original ice thickness of the endpoint, the change of the original ice thickness of the endpoint is intuitively reflected, thereby ensuring the accuracy of the determined current ice state.

[0170] Optionally, in order to ensure the accuracy of the prediction of ice thickness of overhead lines, in one embodiment, the ice growth dynamics theory is introduced to adjust and integrate the predicted ice thickness predicted by the thickness prediction model. Therefore, Figure 5 As shown, a method for adjusting the output result of the thickness prediction model is provided, which specifically includes the following steps:

[0171] S501: Input the multi-scale time characteristics into a thickness prediction model corresponding to the current ice cover state to obtain an initial predicted thickness.

[0172] Optionally, the multi-scale time characteristics can be input into a thickness prediction model corresponding to the current ice cover state, so that the thickness prediction model analyzes and calculates the multi-scale time characteristics based on the model parameters to obtain an initial predicted thickness output by the thickness prediction model. In the embodiment of the present application, the initial predicted thickness output by the thickness prediction model is:

[0173] (30)

[0174] S502: Adjust the initial predicted ice thickness according to an ice adjustment parameter representing an ice growth law of a current ice state to obtain a predicted ice thickness of the overhead line in a future period.

[0175] The ice adjustment parameter is a parameter used to adjust the initial predicted thickness.

[0176] First, based on the current icing state, it is necessary to determine an icing adjustment parameter that characterizes the ice growth pattern of the current icing state. It should be noted that each candidate icing state corresponds to an icing adjustment parameter. It is understood that since ice thickness does not need to be calculated for the ice-free state, in this embodiment of the present application, only icing adjustment parameters for candidate icing states other than the ice-free state are considered.

[0177] Optionally, for the initial icing state and the stable growth state, the same icing adjustment parameter can be expressed as:

[0178]

[0179] in, is the growth adjustment factor; is the ice growth function, Expressed as:

[0180] (31)

[0181] in, is the growth coefficient.

[0182] For the melting state, the corresponding ice adjustment parameters can be expressed as:

[0183]

[0184] in, is the melting adjustment factor; is the ice melting function, Expressed as:

[0185] (32)

[0186] in, is the melting coefficient.

[0187] Furthermore, the corresponding ice adjustment parameters are calculated according to the current ice state, and the initial ice thickness is adjusted according to the ice adjustment parameters. For example, if the current ice state is the initial ice state or the stable growth state, the predicted ice thickness can be expressed as:

[0188] (33)

[0189] If the current ice cover state is melting, the predicted ice cover thickness can be expressed as:

[0190] (34)

[0191] In this embodiment, by introducing ice adjustment parameters, that is, combining the dynamic equations of ice growth and melting processes in the physical model with the thickness prediction model, the prediction results are adjusted through state-specific growth / melting coefficients to make the prediction more consistent with the physical laws of ice formation.

[0192] Optionally, in order to ensure the prediction accuracy of the thickness prediction model, in one embodiment, as Figure 6 As shown, a method for adjusting the thickness prediction model is provided, which specifically includes the following steps:

[0193] S601: Obtain the actual ice thickness of the overhead lines in the future period.

[0194] The actual ice thickness is the ice thickness of the overhead lines measured in the future period.

[0195] Optionally, in the future, the ice thickness at the monitoring point of the overhead line can be measured by manual measurement or automatic machine measurement to serve as the actual ice thickness.

[0196] S602: Determine a prediction error of a thickness prediction model corresponding to a current ice cover state according to the actual ice cover thickness and the predicted ice cover thickness.

[0197] Optionally, the difference between the actual ice thickness and the predicted ice thickness can be used as the prediction error of the thickness prediction model corresponding to the current ice state. Specifically, the prediction error can be calculated using the following formula (35):

[0198] (35)

[0199] in, is the actual ice thickness; To predict ice thickness; is the prediction error.

[0200] S603: Adjust the thickness prediction model corresponding to the current ice cover state according to the prediction error.

[0201] In this embodiment, the thickness prediction model is adjusted by introducing a prediction error, thereby ensuring that the prediction accuracy of the thickness prediction model is improved.

[0202] It should be noted that during the thickness prediction process, if it is detected that the ice cover state of the overhead line may change within the target period, in order to ensure the accuracy of the prediction results, a state mixed prediction is required. First, the state transition probability can be expressed by the following formula (36):

[0203] (36)

[0204] Furthermore, if the ice cover state is determined to have changed based on the state transition probability, the hybrid prediction process can be expressed by the following formula (37):

[0205] (37)

[0206] in, is the status In time The prediction weights ensure .

[0207] Optionally, in order to verify the effectiveness of the method for predicting the ice thickness of overhead lines provided in the embodiments of the present application, in one embodiment, three monitoring points of a 330kV transmission line in a mountainous area were selected for experimental verification. The line spans multiple micro-topographic areas, and the icing conditions are complex and changeable. The experimental data includes real-time monitoring data from November 2021 to March 2022, with a sampling interval of 1 hour and a total of 3624 time points. The data contained in each monitoring point is shown in Table 1.

[0208] The dataset is divided into training, validation, and test sets in a ratio of 7:2:1. Table 2 below shows some examples of the original data.

[0209] To verify the effectiveness of the proposed method, we compared it with traditional ELM, BP neural network, LSTM, and Hadoop-based parallel ELM. All models used the same training and test data. The results are shown in Table 3.

[0210] Table 1 Sample data description

[0211]

[0212] Table 2 Partial real-time monitoring data of monitoring point P1

[0213]

[0214] Table 3 Comparison of prediction accuracy of different methods (test set)

[0215]

[0216] Figure 7 FIG1 is a flow chart of a method for predicting ice thickness of overhead lines in another embodiment. Based on the above embodiment, this embodiment provides an optional example of a method for predicting ice thickness of overhead lines. Figure 7 The specific implementation process is as follows:

[0217] S701, obtaining time-series monitoring data at monitoring points on overhead lines within a target period.

[0218] Among them, the time series monitoring data includes meteorological monitoring data and line status monitoring data at each sampling time within the target period.

[0219] S702 , extracting time series monitoring data using time windows of different lengths to obtain data sequences of different lengths.

[0220] S703 , performing statistical analysis on the meteorological monitoring data and line status monitoring data at the sampling time contained in each data sequence to obtain statistical parameters corresponding to each data sequence.

[0221] S704: Determine the multi-scale time characteristics of the time series monitoring data according to the statistical parameters corresponding to each data sequence.

[0222] S705 , determining a weighted difference of the original ice thickness at the endpoints according to the original ice thickness at the endpoints in the line status monitoring data at each sampling moment within the target period.

[0223] S706 : Select the current icing state of the overhead line from the candidate icing states according to the weighted difference and the meteorological monitoring data at the last sampling moment.

[0224] Each candidate icing state corresponds to a thickness prediction model. Different thickness prediction models use different initialization strategies and activation functions during training. The initialization strategy and activation function used by any thickness prediction model during training are determined based on the candidate icing state corresponding to the thickness prediction model.

[0225] Optionally, whether the overhead line meets the ice growth conditions is determined based on the temperature and humidity at the end sampling time within the target time period; if the original ice thickness at the endpoint at the end sampling time is less than the initial ice thickness threshold, the weighted difference is greater than zero, and the overhead line meets the ice growth conditions, then the current ice state of the overhead line is determined to be the initial ice state; if the original ice thickness at the endpoint at the end sampling time is greater than or equal to the initial ice thickness threshold, the weighted difference is greater than zero, and the overhead line meets the ice growth conditions, then the current ice state of the overhead line is determined to be a stable growth state; if the weighted difference is less than zero, and the overhead line does not meet the ice growth conditions, then the current ice state of the overhead line is determined to be a melting state; if the original ice thickness at the endpoint is less than the minimum detectable ice thickness, and the overhead line does not meet the ice growth conditions, then the current ice state of the overhead line is determined to be an ice-free state.

[0226] S707: Input the multi-scale time characteristics into a thickness prediction model corresponding to the current ice cover state to obtain an initial predicted thickness.

[0227] S708: Adjust the initial predicted ice thickness according to an ice adjustment parameter representing an ice growth law of a current ice state to obtain a predicted ice thickness of the overhead line in a future period.

[0228] S709: Obtain the actual ice thickness of the overhead lines in the future period.

[0229] S710: Determine a prediction error of a thickness prediction model corresponding to a current ice cover state according to the actual ice cover thickness and the predicted ice cover thickness.

[0230] S711: Adjust the thickness prediction model corresponding to the current ice cover state according to the prediction error.

[0231] The specific process of S701-S711 can be found in the description of the above method embodiment. The implementation principle and technical effects are similar and will not be repeated here.

[0232] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0233] Based on the same inventive concept, embodiments of the present application also provide an overhead line ice thickness prediction device for implementing the aforementioned overhead line ice thickness prediction method. The solution provided by this device is similar to the solution described in the aforementioned method. Therefore, the specific limitations of one or more embodiments of the overhead line ice thickness prediction device provided below can be found in the aforementioned definitions of the overhead line ice thickness prediction method and will not be further elaborated here.

[0234] In an exemplary embodiment, Figure 8 As shown, an overhead line ice thickness prediction device 800 is provided, comprising: a data acquisition module 810, a feature extraction module 820, a state selection module 830 and a thickness prediction module 840, wherein:

[0235] The data acquisition module 810 is used to acquire time series monitoring data at monitoring points on the overhead line within the target period; wherein the time series monitoring data includes meteorological monitoring data and line status monitoring data at each sampling moment within the target period.

[0236] The feature extraction module 820 is used to extract multi-scale time features of time series monitoring data using time windows of different lengths.

[0237] The state selection module 830 is used to select the current icing state of the overhead line from the candidate icing states according to the time series monitoring data; wherein each candidate icing state corresponds to a thickness prediction model.

[0238] The thickness prediction module 840 is used to determine the predicted ice thickness of the overhead line in the future period based on the thickness prediction model corresponding to the current ice state and according to the multi-scale time characteristics.

[0239] The above-mentioned overhead line ice thickness prediction device, after obtaining the time series monitoring data of meteorological monitoring data and line status monitoring data at the monitoring points on the overhead line within the target period, including the meteorological monitoring data and line status monitoring data at each sampling time within the target period, uses time windows of different lengths to extract the multi-scale time characteristics of the time series monitoring data, comprehensively considers the characteristics of different scales and multiple dimensions, effectively captures the law of change of meteorological parameters over time, and provides key data support for the dynamic prediction of the icing process; further, according to the time series monitoring data, the current icing state of the overhead line is selected from the candidate icing states, and each candidate icing state corresponds to a thickness prediction model. By formulating targeted thickness prediction models for different icing states, accurate prediction of different icing stages is achieved; finally, based on the thickness prediction model corresponding to the current icing state, according to the multi-scale time characteristics, the predicted icing thickness of the overhead line in the future period is determined, thereby ensuring the accuracy of the determined predicted icing thickness.

[0240] In one embodiment, the feature extraction module 820 is specifically configured to:

[0241] Time windows of different lengths are used to extract time series monitoring data to obtain data sequences of different lengths. For each length of data sequence, the meteorological monitoring data and line status monitoring data at the sampling moment contained in the data sequence are statistically analyzed to obtain the statistical parameters corresponding to the data sequence. Based on the statistical parameters corresponding to each data sequence, the multi-scale time characteristics of the time series monitoring data are determined.

[0242] In one embodiment, the line status monitoring data includes at least the original ice thickness at the endpoint of the overhead line; the status selection module 830 includes:

[0243] The difference determination unit is used to determine the weighted difference of the original ice thickness of the endpoint according to the original ice thickness of the endpoint at each sampling time within the target time period.

[0244] The state selection unit is used to select the current icing state of the overhead line from the candidate icing states according to the weighted difference and the meteorological monitoring data at the end sampling time in the target period.

[0245] In one embodiment, the candidate icing states include an ice-free state, an initial icing state, a stable growth state, and a melting state; the meteorological monitoring data includes at least the temperature and humidity of the environment in which the overhead line is located; and the state selection unit is specifically configured to:

[0246] Based on the temperature and humidity at the last sampling time within the target period, determine whether the overhead line meets the ice growth conditions; if the original ice thickness at the endpoint at the last sampling time is less than the initial ice thickness threshold, the weighted difference is greater than zero, and the overhead line meets the ice growth conditions, then the current ice state of the overhead line is determined to be the initial ice state; if the original ice thickness at the endpoint at the last sampling time is greater than or equal to the initial ice thickness threshold, the weighted difference is greater than zero, and the overhead line meets the ice growth conditions, then the current ice state of the overhead line is determined to be the stable growth state; if the weighted difference is less than zero, and the overhead line does not meet the ice growth conditions, then the current ice state of the overhead line is determined to be the melting state; if the original ice thickness at the endpoint is less than the minimum detectable ice thickness, and the overhead line does not meet the ice growth conditions, then the current ice state of the overhead line is determined to be the ice-free state.

[0247] In one embodiment, the thickness prediction module 840 is specifically configured to:

[0248] The multi-scale time characteristics are input into the thickness prediction model corresponding to the current icing state to obtain the initial predicted thickness. The initial predicted thickness is adjusted according to the icing adjustment parameters that characterize the icing growth law of the current icing state to obtain the predicted icing thickness of the overhead line in the future period.

[0249] In one embodiment, the apparatus 800 for predicting ice thickness of overhead lines is further configured to:

[0250] The actual ice thickness of the overhead lines in the future period is obtained; based on the actual ice thickness and the predicted ice thickness, the prediction error of the thickness prediction model corresponding to the current ice state is determined; and based on the prediction error, the thickness prediction model corresponding to the current ice state is adjusted.

[0251] In one embodiment, different thickness prediction models use different initialization strategies and activation functions during training, and the initialization strategy and activation function used by any thickness prediction model during training are determined according to the candidate icing state corresponding to the thickness prediction model.

[0252] Each module in the above-mentioned device for predicting ice thickness on overhead lines can be implemented in whole or in part through software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in hardware form, or can be stored in a memory in the computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0253] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Figure 9 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O) and a communication interface. The processor, memory and input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for predicting the ice thickness of an overhead line is implemented.

[0254] Those skilled in the art will understand that Figure 9 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0255] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:

[0256] Acquire time series monitoring data at monitoring points on the overhead line within a target period; wherein the time series monitoring data includes meteorological monitoring data and line status monitoring data at each sampling time within the target period;

[0257] Use time windows of different lengths to extract multi-scale temporal features of time series monitoring data;

[0258] According to the time series monitoring data, the current icing state of the overhead line is selected from the candidate icing states; each candidate icing state corresponds to a thickness prediction model;

[0259] Based on the thickness prediction model corresponding to the current icing state and the multi-scale time characteristics, the predicted icing thickness of the overhead lines in the future period is determined.

[0260] In one embodiment, when the processor executes the computer program to extract multi-scale time features of time series monitoring data using time windows of different lengths, the processor further implements the following steps:

[0261] Time windows of different lengths are used to extract time series monitoring data to obtain data sequences of different lengths. For each length of data sequence, the meteorological monitoring data and line status monitoring data at the sampling moment contained in the data sequence are statistically analyzed to obtain the statistical parameters corresponding to the data sequence. Based on the statistical parameters corresponding to each data sequence, the multi-scale time characteristics of the time series monitoring data are determined.

[0262] In one embodiment, the line state monitoring data includes at least the original ice thickness at the endpoint of the overhead line; when the processor executes the computer program to select the current ice state of the overhead line from the candidate ice states based on the time-series monitoring data, the processor further implements the following steps:

[0263] According to the original ice thickness of the endpoints at each sampling time within the target period, the weighted difference of the original ice thickness of the endpoints is determined; according to the weighted difference and the meteorological monitoring data at the last sampling time within the target period, the current ice state of the overhead line is selected from the candidate ice state.

[0264] In one embodiment, the candidate icing states include an ice-free state, an initial icing state, a stable growth state, and a melting state; the meteorological monitoring data includes at least the temperature and humidity of the environment in which the overhead line is located; and the processor executing the computer program selects the current icing state of the overhead line from the candidate icing states based on the weighted difference and the meteorological monitoring data at the end sampling time within the target period, and further implements the following steps:

[0265] Based on the temperature and humidity at the last sampling time within the target period, determine whether the overhead line meets the ice growth conditions; if the original ice thickness at the endpoint at the last sampling time is less than the initial ice thickness threshold, the weighted difference is greater than zero, and the overhead line meets the ice growth conditions, then the current ice state of the overhead line is determined to be the initial ice state; if the original ice thickness at the endpoint at the last sampling time is greater than or equal to the initial ice thickness threshold, the weighted difference is greater than zero, and the overhead line meets the ice growth conditions, then the current ice state of the overhead line is determined to be the stable growth state; if the weighted difference is less than zero, and the overhead line does not meet the ice growth conditions, then the current ice state of the overhead line is determined to be the melting state; if the original ice thickness at the endpoint is less than the minimum detectable ice thickness, and the overhead line does not meet the ice growth conditions, then the current ice state of the overhead line is determined to be the ice-free state.

[0266] In one embodiment, when a processor executes a computer program to determine a predicted ice thickness of an overhead line in a future time period based on a thickness prediction model corresponding to a current ice state and according to multi-scale time characteristics, the processor further implements the following steps:

[0267] The multi-scale time characteristics are input into the thickness prediction model corresponding to the current icing state to obtain the initial predicted thickness. The initial predicted thickness is adjusted according to the icing adjustment parameters that characterize the icing growth law of the current icing state to obtain the predicted icing thickness of the overhead line in the future period.

[0268] In one embodiment, when the processor executes the computer program, it further implements the following steps:

[0269] The actual ice thickness of the overhead lines in the future period is obtained; based on the actual ice thickness and the predicted ice thickness, the prediction error of the thickness prediction model corresponding to the current ice state is determined; and based on the prediction error, the thickness prediction model corresponding to the current ice state is adjusted.

[0270] In one embodiment, different thickness prediction models use different initialization strategies and activation functions during training, and the initialization strategy and activation function used by any thickness prediction model during training are determined according to the candidate icing state corresponding to the thickness prediction model.

[0271] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0272] Acquire time series monitoring data at monitoring points on the overhead line within a target period; wherein the time series monitoring data includes meteorological monitoring data and line status monitoring data at each sampling time within the target period;

[0273] Use time windows of different lengths to extract multi-scale temporal features of time series monitoring data;

[0274] According to the time series monitoring data, the current icing state of the overhead line is selected from the candidate icing states; each candidate icing state corresponds to a thickness prediction model;

[0275] Based on the thickness prediction model corresponding to the current icing state and the multi-scale time characteristics, the predicted icing thickness of the overhead lines in the future period is determined.

[0276] In one embodiment, when the processor executes the computer program to extract multi-scale time features of time series monitoring data using time windows of different lengths, the processor further implements the following steps:

[0277] Time windows of different lengths are used to extract time series monitoring data to obtain data sequences of different lengths. For each length of data sequence, the meteorological monitoring data and line status monitoring data at the sampling moment contained in the data sequence are statistically analyzed to obtain the statistical parameters corresponding to the data sequence. Based on the statistical parameters corresponding to each data sequence, the multi-scale time characteristics of the time series monitoring data are determined.

[0278] In one embodiment, the line state monitoring data includes at least the original ice thickness at the endpoint of the overhead line; when the processor executes the computer program to select the current ice state of the overhead line from the candidate ice states based on the time-series monitoring data, the processor further implements the following steps:

[0279] According to the original ice thickness of the endpoints at each sampling time within the target period, the weighted difference of the original ice thickness of the endpoints is determined; according to the weighted difference and the meteorological monitoring data at the last sampling time within the target period, the current ice state of the overhead line is selected from the candidate ice state.

[0280] In one embodiment, the candidate icing states include an ice-free state, an initial icing state, a stable growth state, and a melting state; the meteorological monitoring data includes at least the temperature and humidity of the environment in which the overhead line is located; and the processor executing the computer program selects the current icing state of the overhead line from the candidate icing states based on the weighted difference and the meteorological monitoring data at the end sampling time within the target period, and further implements the following steps:

[0281] Based on the temperature and humidity at the last sampling time within the target period, determine whether the overhead line meets the ice growth conditions; if the original ice thickness at the endpoint at the last sampling time is less than the initial ice thickness threshold, the weighted difference is greater than zero, and the overhead line meets the ice growth conditions, then the current ice state of the overhead line is determined to be the initial ice state; if the original ice thickness at the endpoint at the last sampling time is greater than or equal to the initial ice thickness threshold, the weighted difference is greater than zero, and the overhead line meets the ice growth conditions, then the current ice state of the overhead line is determined to be the stable growth state; if the weighted difference is less than zero, and the overhead line does not meet the ice growth conditions, then the current ice state of the overhead line is determined to be the melting state; if the original ice thickness at the endpoint is less than the minimum detectable ice thickness, and the overhead line does not meet the ice growth conditions, then the current ice state of the overhead line is determined to be the ice-free state.

[0282] In one embodiment, when a processor executes a computer program to determine a predicted ice thickness of an overhead line in a future time period based on a thickness prediction model corresponding to a current ice state and according to multi-scale time characteristics, the processor further implements the following steps:

[0283] The multi-scale time characteristics are input into the thickness prediction model corresponding to the current icing state to obtain the initial predicted thickness. The initial predicted thickness is adjusted according to the icing adjustment parameters that characterize the icing growth law of the current icing state to obtain the predicted icing thickness of the overhead line in the future period.

[0284] In one embodiment, when the processor executes the computer program, it further implements the following steps:

[0285] The actual ice thickness of the overhead lines in the future period is obtained; based on the actual ice thickness and the predicted ice thickness, the prediction error of the thickness prediction model corresponding to the current ice state is determined; and based on the prediction error, the thickness prediction model corresponding to the current ice state is adjusted.

[0286] In one embodiment, different thickness prediction models use different initialization strategies and activation functions during training, and the initialization strategy and activation function used by any thickness prediction model during training are determined according to the candidate icing state corresponding to the thickness prediction model.

[0287] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps:

[0288] Acquire time series monitoring data at monitoring points on the overhead line within a target period; wherein the time series monitoring data includes meteorological monitoring data and line status monitoring data at each sampling time within the target period;

[0289] Use time windows of different lengths to extract multi-scale temporal features of time series monitoring data;

[0290] According to the time series monitoring data, the current icing state of the overhead line is selected from the candidate icing states; each candidate icing state corresponds to a thickness prediction model;

[0291] Based on the thickness prediction model corresponding to the current icing state and the multi-scale time characteristics, the predicted icing thickness of the overhead lines in the future period is determined.

[0292] In one embodiment, when the processor executes the computer program to extract multi-scale time features of time series monitoring data using time windows of different lengths, the processor further implements the following steps:

[0293] Time windows of different lengths are used to extract time series monitoring data to obtain data sequences of different lengths. For each length of data sequence, the meteorological monitoring data and line status monitoring data at the sampling moment contained in the data sequence are statistically analyzed to obtain the statistical parameters corresponding to the data sequence. Based on the statistical parameters corresponding to each data sequence, the multi-scale time characteristics of the time series monitoring data are determined.

[0294] In one embodiment, the line state monitoring data includes at least the original ice thickness at the endpoint of the overhead line; when the processor executes the computer program to select the current ice state of the overhead line from the candidate ice states based on the time-series monitoring data, the processor further implements the following steps:

[0295] According to the original ice thickness of the endpoints at each sampling time within the target period, the weighted difference of the original ice thickness of the endpoints is determined; according to the weighted difference and the meteorological monitoring data at the last sampling time within the target period, the current ice state of the overhead line is selected from the candidate ice state.

[0296] In one embodiment, the candidate icing states include an ice-free state, an initial icing state, a stable growth state, and a melting state; the meteorological monitoring data includes at least the temperature and humidity of the environment in which the overhead line is located; and the processor executing the computer program selects the current icing state of the overhead line from the candidate icing states based on the weighted difference and the meteorological monitoring data at the end sampling time within the target period, and further implements the following steps:

[0297] Based on the temperature and humidity at the last sampling time within the target period, determine whether the overhead line meets the ice growth conditions; if the original ice thickness at the endpoint at the last sampling time is less than the initial ice thickness threshold, the weighted difference is greater than zero, and the overhead line meets the ice growth conditions, then the current ice state of the overhead line is determined to be the initial ice state; if the original ice thickness at the endpoint at the last sampling time is greater than or equal to the initial ice thickness threshold, the weighted difference is greater than zero, and the overhead line meets the ice growth conditions, then the current ice state of the overhead line is determined to be the stable growth state; if the weighted difference is less than zero, and the overhead line does not meet the ice growth conditions, then the current ice state of the overhead line is determined to be the melting state; if the original ice thickness at the endpoint is less than the minimum detectable ice thickness, and the overhead line does not meet the ice growth conditions, then the current ice state of the overhead line is determined to be the ice-free state.

[0298] In one embodiment, when a processor executes a computer program to determine a predicted ice thickness of an overhead line in a future time period based on a thickness prediction model corresponding to a current ice state and according to multi-scale time characteristics, the processor further implements the following steps:

[0299] The multi-scale time characteristics are input into the thickness prediction model corresponding to the current icing state to obtain the initial predicted thickness. The initial predicted thickness is adjusted according to the icing adjustment parameters that characterize the icing growth law of the current icing state to obtain the predicted icing thickness of the overhead line in the future period.

[0300] In one embodiment, when the processor executes the computer program, it further implements the following steps:

[0301] The actual ice thickness of the overhead lines in the future period is obtained; based on the actual ice thickness and the predicted ice thickness, the prediction error of the thickness prediction model corresponding to the current ice state is determined; and based on the prediction error, the thickness prediction model corresponding to the current ice state is adjusted.

[0302] In one embodiment, different thickness prediction models use different initialization strategies and activation functions during training, and the initialization strategy and activation function used by any thickness prediction model during training are determined according to the candidate icing state corresponding to the thickness prediction model.

[0303] It should be noted that the data involved in this application (including but not limited to time-series monitoring data at monitoring points on overhead lines during the target period, etc.) are information and data that have been fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0304] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the above-mentioned embodiments. In particular, any reference to memory, database, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The databases involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processors involved in the various embodiments provided herein may be, but are not limited to, general-purpose processors, central processing units (CPUs), graphics processing units (GPUs), digital signal processors (DSPs), programmable logic devices (PLDs), data processing logic devices based on quantum computing, and the like.

[0305] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0306] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A method for predicting the icing thickness of an overhead line, characterized in that The method includes: Obtaining time-series monitoring data at a monitoring point on an overhead line during a target period; wherein, the time-series monitoring data includes meteorological monitoring data and line state monitoring data at each sampling moment during the target period; Using time windows of different lengths to extract multi-scale time features of the time-series monitoring data; Selecting the current icing state of the overhead line from candidate icing states according to the time-series monitoring data; wherein, each candidate icing state corresponds to a thickness prediction model; Based on the thickness prediction model corresponding to the current icing state, determining the predicted icing thickness of the overhead line in a future period according to the multi-scale time features.

2. The method according to claim 1, characterized in that The step of using time windows of different lengths to extract multi-scale time features of the time-series monitoring data includes: Using time windows of different lengths to extract the time-series monitoring data to obtain data sequences of different lengths; For each data sequence of a length, performing statistical analysis on the meteorological monitoring data and line state monitoring data at the sampling moments included in the data sequence to obtain statistical parameters corresponding to the data sequence; Determining the multi-scale time features of the time-series monitoring data according to the statistical parameters corresponding to the data sequences.

3. The method according to claim 1, characterized in that, The line state monitoring data at least includes the original icing thickness at the endpoints of the overhead line; The step of selecting the current icing state of the overhead line from candidate icing states according to the time-series monitoring data includes: Determining the weighted difference of the original icing thickness at the endpoints according to the original icing thickness at the endpoints at each sampling moment during the target period; Selecting the current icing state of the overhead line from candidate icing states according to the weighted difference and the meteorological monitoring data at the last sampling moment during the target period.

4. The method according to claim 3, characterized in that The candidate icing states include a non-icing state, an initial icing state, a stable growth state, and a melting state; the meteorological monitoring data at least includes the temperature and humidity of the environment where the overhead line is located; The step of selecting the current icing state of the overhead line from candidate icing states according to the weighted difference and the meteorological monitoring data at the last sampling moment during the target period includes: Determining whether the overhead line meets the icing growth condition according to the temperature and humidity at the last sampling moment during the target period; If the original icing thickness at the endpoints at the last sampling moment is less than the initial icing thickness threshold, the weighted difference is greater than zero, and the overhead line meets the icing growth condition, then determining the current icing state of the overhead line as the initial icing state; If the original icing thickness at the endpoints at the last sampling moment is greater than or equal to the initial icing thickness threshold, the weighted difference is greater than zero, and the overhead line meets the icing growth condition, then determining the current icing state of the overhead line as the stable growth state; If the weighted difference is less than zero, and the overhead line does not meet the icing growth condition, then determining the current icing state of the overhead line as the melting state; If the original icing thickness at the endpoints is less than the minimum detectable icing thickness, and the overhead line does not meet the icing growth condition, then determining the current icing state of the overhead line as the non-icing state.

5. The method according to claim 1, characterized in that, Based on the thickness prediction model corresponding to the current ice-covered state, determining the predicted ice-covered thickness of the overhead line in a future period according to the multi-scale time features includes: Inputting the multi-scale time features into the thickness prediction model corresponding to the current ice-covered state to obtain an initial predicted thickness; Adjusting the initial predicted thickness according to the ice-covered adjustment parameter characterizing the ice-covered growth law of the current ice-covered state to obtain the predicted ice-covered thickness of the overhead line in a future period.

6. The method according to claim 1, wherein The method further includes: Obtaining the actual ice-covered thickness of the overhead line in a future period; Determining the prediction error of the thickness prediction model corresponding to the current ice-covered state according to the actual ice-covered thickness and the predicted ice-covered thickness; Adjusting the thickness prediction model corresponding to the current ice-covered state according to the prediction error.

7. The method according to any one of claims 1-6, characterized in that, The initialization strategies and activation functions adopted by different thickness prediction models during training are different, and the initialization strategy and activation function adopted by any thickness prediction model during training are determined according to the candidate ice-covered state corresponding to the thickness prediction model.

8. An icing thickness prediction device for an overhead line, characterized in that, The device includes: A data acquisition module, configured to acquire time-series monitoring data at a monitoring point on the overhead line during a target period; wherein, the time-series monitoring data includes meteorological monitoring data and line state monitoring data at each sampling moment during the target period; A feature extraction module, configured to extract multi-scale time features of the time-series monitoring data by using time windows of different lengths; A state selection module, configured to select the current ice-covered state of the overhead line from candidate ice-covered states according to the time-series monitoring data; wherein, each candidate ice-covered state corresponds to a thickness prediction model; A thickness prediction module, configured to determine the predicted ice-covered thickness of the overhead line in a future period based on the thickness prediction model corresponding to the current ice-covered state according to the multi-scale time features.

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

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images