Power transmission line icing disaster detection method and device, electronic equipment and storage medium

Abstract

Embodiments of the present application disclose a power transmission line icing disaster detection method and device, electronic equipment and storage medium. The method comprises: fusing an ALD model by using a KDLV algorithm, extracting dynamic and static characteristics of data, and further detecting an icing disaster of a power transmission line by using an autoregressive model in the KDLV algorithm. A T statistical limit of a first autoregressive model constructed based on training data can represent a relatively normal (without icing disaster) power transmission line condition. A T statistical value of a second autoregressive model constructed based on test data can be used to represent an icing disaster condition of a power transmission line to be tested. Therefore, by comparing the T statistical limit corresponding to the training data and the T statistical value corresponding to the test data, a detection result of an extreme icing disaster of the power transmission line can be obtained. On the premise of using a small amount of icing weight data as a training set, the early warning of the extreme icing disaster of the power transmission line of the power grid system is adaptively completed, and the accuracy of the icing disaster detection is improved.

Description

Technical Field

[0001] This application relates to the field of power transmission line inspection technology, specifically to methods, devices, electronic equipment, and storage media for detecting freezing disasters on power transmission lines. Background Technology

[0002] The power system plays a vital role in modern society and daily life. It is not only the cornerstone of industrial production but also a basic necessity for residents' daily lives. A stable power supply directly impacts stable economic growth and the overall functioning of society. However, the stable operation of the power system is closely related to weather conditions. Transmission lines are exposed to the elements year-round, constantly subjected to wind, sun, and even extreme weather conditions such as rain and snow, making them highly susceptible to safety problems. Therefore, it is necessary to conduct ice storm detection on transmission lines.

[0003] With the continuous improvement of machine learning and data acquisition capabilities, data-driven methods have gradually become the mainstream for icing detection. However, these methods require a large number of fault samples (samples of icing disasters) to better identify the icing state when training the model. However, when facing a once-in-a-century extreme freezing disaster, it is difficult to obtain extremely low-probability samples, so the above methods will have difficulties in training the model, thus affecting the final icing disaster detection rate. Summary of the Invention

[0004] This application discloses a method, device, electronic equipment, and storage medium for detecting icing disasters on power transmission lines. It can adaptively complete the early warning of extreme icing disasters on power grid transmission lines using a small amount of icing weight data as a training set, thereby improving the accuracy of icing disaster detection.

[0005] This application discloses a method for detecting freezing disasters on power transmission lines, the method comprising:

[0006] Obtain the first dataset; the first dataset includes icing weight data of transmission lines and meteorological condition data; preprocess the first dataset to obtain the second dataset;

[0007] The second dataset is projected using the kernel principal component extraction strategy of the kernel dynamic latent variable (KDLV) algorithm to obtain the principal component matrix;

[0008] Training data and test data are extracted from the principal component matrix; wherein, the training data is the data in the principal component matrix representing the transmission line in a normal state and in a first icing state; the test data is the data in the principal component matrix representing the transmission line in a normal state and in a second icing state; the second icing state indicates a higher degree of icing than the first icing state;

[0009] The first autoregressive model is constructed using the dynamic components in the training data through the KDLV algorithm, and the T-statistic limit corresponding to the first autoregressive model is calculated.

[0010] Calculate the ALD value of each data point in the test data, and filter out the first test sub-data and the second test sub-data from the test data based on the ALD value;

[0011] The first autoregressive model is updated using the first test sub-data to update the T statistical limit;

[0012] Using the KDLV algorithm, a second autoregressive model is constructed using the dynamic components in the second test sub-data, and the T-statistic corresponding to the second autoregressive model is calculated.

[0013] The detection results of the transmission line freezing disaster are determined based on the comparison between the T statistical limit and the T statistical value.

[0014] As an optional implementation, determining the detection result of transmission line freezing disaster based on the comparison result of the T statistical limit and the T statistical value includes:

[0015] If the T statistical value is greater than the T statistical limit, an alarm is output; the alarm is used to indicate that there is freezing damage to the transmission line; or,

[0016] If the T-statistic is less than or equal to the T-statistic, then it is determined that the transmission line does not suffer from freezing disaster.

[0017] As an optional implementation, the step of filtering the first test sub-data and the second test sub-data from the test data based on the ALD value includes:

[0018] Based on the comparison results between the ALD value and the error threshold, the first test sub-data and the second test sub-data are selected from the test data;

[0019] The first test sub-data is the test data whose corresponding ALD value is greater than the error threshold; the second test sub-data is the test data whose corresponding ALD value is greater than or equal to the error threshold.

[0020] As an optional implementation, both the first autoregressive model and the second autoregressive model are AR(2) models; there are at least three second test data points between every two adjacent first test data points.

[0021] As an optional implementation, the preprocessing of the first dataset to obtain the second dataset includes:

[0022] The data in the first dataset is decomposed into K intrinsic mode functions using the Variational Mode Decomposition (VMD) algorithm to perform noise reduction on the first dataset; K is a positive integer greater than or equal to 1.

[0023] The time-domain and frequency-domain features of the K intrinsic mode functions are extracted, and the extracted time-domain and frequency-domain features are standardized to obtain the second dataset.

[0024] As an optional implementation, the meteorological condition data includes at least the average sunshine intensity, which is the average of the sunshine intensity values ​​at N sampling times; N is a positive integer greater than or equal to 2.

[0025] As an optional implementation, the meteorological condition data may include at least temperature data in the range of -15°C to 0°C; or, the meteorological condition data may include at least humidity data greater than 80%.

[0026] This application discloses a transmission line freezing disaster detection device, comprising:

[0027] The acquisition module is used to acquire a first dataset; the first dataset includes icing weight data of transmission lines and meteorological condition data;

[0028] The preprocessing module is used to preprocess the first dataset to obtain the second dataset;

[0029] The first processing module is used to project the second dataset using the kernel principal component extraction strategy of the kernel dynamic latent variable (KDLV) algorithm to obtain the principal component matrix;

[0030] An extraction module is used to extract training data and test data from the principal component matrix; wherein, the training data is data in the principal component matrix representing the transmission line in a normal state and in a first icing state; the test data is data in the principal component matrix representing the transmission line in a normal state and in a second icing state; the second icing state indicates a higher degree of icing than the first icing state;

[0031] The second processing module is used to build a first autoregressive model using the dynamic components in the training data through the KDLV algorithm, and to calculate the T-statistic limit corresponding to the first autoregressive model.

[0032] The calculation module is used to calculate the ALD value of each data in the test data, and to filter out the first test sub-data and the second test sub-data from the test data based on the ALD value;

[0033] The update module is used to update the first autoregressive model using the first test sub-data in order to update the T statistical limit;

[0034] The third processing module is used to build a second autoregressive model using the dynamic components in the second test sub-data through the KDLV algorithm, and to calculate the T-statistic corresponding to the second autoregressive model.

[0035] The determination module is used to determine the detection results of ice disasters on transmission lines based on the comparison results of the T statistical limit and the T statistical value.

[0036] This application discloses an electronic device, which includes a memory and a processor. The memory stores a computer program, and when the computer program is executed by the processor, the processor implements any of the methods for detecting freezing of power transmission lines disclosed in this application.

[0037] This application discloses a computer-readable storage medium storing a computer program thereon, characterized in that the computer program, when executed by a processor, implements any of the methods for detecting freezing of transmission lines disclosed in this application.

[0038] Compared with related technologies, the embodiments of this application have the following beneficial effects:

[0039] In this embodiment, by collecting a small amount of icing weight data and meteorological condition data, and fusing the KDLV algorithm with the ALD model, dynamic and static features of the data can be extracted. This allows for the detection of icing disasters on transmission lines based on the autoregressive model of the KDLV algorithm. Specifically, the T-statistic limit of the first autoregressive model constructed based on training data characterizes a relatively normal (no icing disaster) transmission line condition; the T-statistic value of the second autoregressive model constructed based on test data can be used to characterize the icing disaster condition of the transmission line under test. Therefore, by comparing the T-statistic limit and the T-statistic value, the detection result of extreme icing disasters on the transmission line can be obtained.

[0040] It is evident that implementing the embodiments of this application can adaptively complete the early warning of extreme freezing disasters on power grid transmission lines using only a small amount of icing weight data as a training set, thereby improving the accuracy of icing disaster detection. Attached Figure Description

[0041] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1 A flowchart illustrating a method for detecting freezing disasters on power transmission lines is disclosed in an embodiment of this application;

[0043] Figure 2 This is a schematic diagram of the structure of a power transmission line freezing disaster detection device disclosed in an embodiment of this application;

[0044] Figure 3 This is a schematic diagram of the structure of an electronic device disclosed in an embodiment of this application. Detailed Implementation

[0045] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0046] It should be noted that the terms "comprising" and "having," and any variations thereof, in the embodiments and accompanying drawings of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the steps or units listed, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.

[0047] This application discloses a method, apparatus, electronic device, and storage medium for detecting icing disasters on power transmission lines. It can adaptively provide early warnings of extreme icing disasters on power grid transmission lines using only a small amount of icing weight data as a training set, thus improving the accuracy of icing disaster detection. These will be described in detail below.

[0048] Please see Figure 1 , Figure 1 This is a flowchart illustrating a method for detecting freezing disasters on power transmission lines disclosed in an embodiment of this application. Figure 1 The method shown can be executed by electronic devices with computing capabilities, such as personal computers and industrial computers. Figure 1 As shown, the method may include the following steps:

[0049] 110. Obtain the first dataset, which includes icing weight data of transmission lines and meteorological condition data.

[0050] In this embodiment, the icing weight data and meteorological condition data of the transmission line can be characteristic data of the transmission line selected from the power grid system at the same sampling time interval. The meteorological condition data may include one or more of the following six meteorological conditions: temperature, humidity, average solar radiation intensity, wind speed, wind direction, and air pressure.

[0051] Here, average solar irradiance refers to the average of the solar irradiance values ​​at N sampling times; N is a positive integer greater than or equal to 2. This average can be an arithmetic mean or a weighted average, without limitation. For example, the average solar irradiance can be calculated using the following formula:

[0052]

[0053] Where a1 a2…a r Q represents the solar radiation intensity values ​​at r sampling times on a given day. Due to the Earth's revolution around the sun, the daily sunshine duration varies before and after the winter solstice, so Q is a balance coefficient used for adjustment.

[0054] It is understandable that instantaneous "solar intensity" only applies to daytime, and the solar intensity value drops rapidly to 0 at night. Therefore, using only the "solar intensity" value as a factor influencing the formation of freezing disasters lacks practicality. Thus, using average solar intensity can more accurately characterize the impact of sunlight on the formation of freezing disasters.

[0055] It should be noted that, except for the average solar radiation intensity, all other meteorological data can be sample values ​​corresponding to the sampling time.

[0056] In addition, in some other possible embodiments, the mechanism of ice formation in the icing mechanism can be combined to further screen meteorological condition data to remove outliers, thereby improving the accuracy of the entire detection algorithm.

[0057] Optionally, temperature data in meteorological condition data can be filtered based on the temperature range of freezing formation in the icing mechanism. For example, temperature data in the range of -15℃ to 0℃ can be retained, while temperature data outside this range can be removed.

[0058] Optionally, humidity data in meteorological condition data can be filtered based on the humidity range formed by temperature in the icing mechanism. For example, this could include retaining humidity data greater than 80% and removing humidity data less than or equal to 80%.

[0059] 120. Preprocess the first dataset to obtain the second dataset.

[0060] In this embodiment, the preprocessing operations on the first dataset may include, but are not limited to, one or more of the following: data cleaning, noise reduction, resampling, type conversion, etc. The purpose of preprocessing is to improve the accuracy of the data, so that the second dataset obtained after preprocessing can more accurately represent the icing state of the transmission line.

[0061] As an optional implementation, the electronic device can perform noise reduction on the first dataset using a VMD algorithm to reduce modal aliasing in the data. Therefore, the electronic device can perform the following steps:

[0062] 1201. Decompose the data in the first dataset into K eigenmode functions using the VMD algorithm, where K is a positive integer greater than or equal to 1. For example, the specific constrained variational process can be shown below:

[0063]

[0064] Among them, {u k} represents the set of variational intrinsic mode function (VIMF) components obtained after decomposition, {ω k} represents the corresponding set of center frequencies. This means taking the partial derivative, where δ(t) represents the unit impulse function, j is the imaginary unit, and f(t) represents the signal before VMD decomposition.

[0065] 1202. Extract the time-domain and frequency features of K intrinsic mode functions, and standardize the extracted time-domain and frequency-domain features to obtain the second dataset.

[0066] The process of standardizing the extracted time-domain features and frequency-domain features can be described as follows:

[0067]

[0068] Where, x k ` represents the standardized input data, x k For the extracted time-domain or frequency-domain features, Let σ represent the mean of the i-th feature. i This represents the standard deviation of the extracted time-domain or frequency-domain features.

[0069] 130. Project the second dataset using the kernel principal component extraction strategy of the kernel dynamic latent variable (KDLV) algorithm to obtain the principal component matrix.

[0070] In this embodiment, the kernel principal component extraction strategy in the KDLV algorithm can be applied to project the data in the second set to obtain the principal component matrix. For example, the principal component matrix can be represented by the following formula:

[0071] X = [x1, ... x2] n ] T (4)

[0072] Furthermore, we can first assume that the matrix of z time delays of the principal component is as shown in the following equation, and this assumption applies to the subsequent step 150:

[0073] X z =[x q-z x q-z+1 …x q-z+n-1 ] T (z=0,...,q-1) (5)

[0074] Where q represents the maximum number of time intervals that can be delayed, and n represents the number of samples.

[0075] 140. Extract training and test data from the principal component matrix.

[0076] In this embodiment, the training data can be data in the principal component matrix representing the transmission line in a normal state and in a first icing state; the test data can be data in the principal component matrix representing the transmission line in a normal state and in a second icing state. The second icing state indicates a higher degree of icing than the first icing state. For example, the second icing state can refer to a more severe icing state, while the first icing state can refer to a slight icing state. The definition of the degree of icing can be set based on business requirements.

[0077] 150. Construct a first autoregressive model using the dynamic components in the training data through the KDLV algorithm, and calculate the T-statistic limit corresponding to the first autoregressive model.

[0078] In step 150, the electronic device may perform the following sub-steps in sequence:

[0079] 1510. Extract dynamic components from training data;

[0080] 1520. Construct the first autoregressive model based on the dynamic components extracted from the training data;

[0081] 1530. Calculate the T-statistic limit corresponding to the first autoregressive model.

[0082] The following section provides a detailed explanation of the three sub-steps mentioned above.

[0083] First, sub-step 1510: extracting dynamic components from the training data can further include the following steps:

[0084] 1511. Obtain the objective function of the KDLV algorithm.

[0085] Suppose the optimization objective of the KDLV algorithm is as follows:

[0086]

[0087] st||w||=1, ||β||=1 (66-2)

[0088] Where β = [β0, ..., β] q-1 ] T X represents z The weights of w, where w is the weight of X. z The weight vector.

[0089] Suppose we use the Kronecker inner product to represent the relationship between β and w, and choose U to represent [X0, X1, ... X q-1 The corresponding objective function can be expressed as follows:

[0090]

[0091] st||w||=1, ||β||=1 (7-2)

[0092] or

[0093]

[0094] st||w||=1, ||β||=1 (8-2)

[0095] Where I represents the identity matrix.

[0096] 1512. Solve the aforementioned objective function using an iterative algorithm.

[0097] In step 1511, it is assumed that w i = [1, 0, ... 0] T The iterative algorithm continues until w i The process ends upon convergence. Use the obtained w... i The dynamic hidden score vector s can be obtained. i and load vector l i The specific formula is as follows:

[0098] s i =X i w i (9)

[0099]

[0100] 1513. When all dynamic components in the feature data have been extracted (assuming there are A dynamic components), then set S = [s1, s2, ..., s...]. A ], L i =[l1 ， l2, ..., l A W = [w1, w2, ..., w] A ]

[0101] and R = W(L) T W) -1 Thus, we can obtain S = XR.

[0102] 1514. Given a new process data x g Its dynamic score s dg and the corresponding residual e g The calculation method is as follows:

[0103]

[0104] Where g represents the dynamic component of the g-th set of feature data.

[0105] In summary, based on the aforementioned steps 1511-1514, the electronic device can extract dynamic components from the training data.

[0106] Then, in step 1520: the first autoregressive model is built based on the dynamic components extracted from the training data. The first autoregressive model can be any type of autoregressive (AR) model, such as a pure autoregressive model, a moving average model, etc.

[0107] Optionally, the first autoregressive model can be a pure autoregressive model, and the model predicts the current value based on the values ​​of the previous two time points. That is, the first autoregressive model can be an AR(2) model. For example, the first autoregressive model can be described using an AR(2) model as follows:

[0108]

[0109] Among them, s` dg This represents the dynamic components after autoregression. and η represents the coefficient value of the autoregression. dg This represents the amount of random error.

[0110] Finally, step 1530: Calculate the T-statistical limit corresponding to the first autoregressive model. The T-statistical limit can refer to the limit calculated using T... 2 Statistical methods are used to determine the statistical limit T of the training data. θ For example, the statistical limit T is T. θ The calculation method is shown in the following formula:

[0111]

[0112] The electronic device can calculate the T-statistic limit of the first autoregressive model corresponding to the training data by performing step 150 as described above.

[0113] 160. Calculate the ALD value of each data point in the test data, and select the first test sub-data and the second test sub-data from the test data based on the ALD value.

[0114] In this embodiment of the application, the test data may be sample data collected within the same time period as the training data; or, the test data may be newly collected sample data after the training data has been selected, as long as the newly collected sample data meets the aforementioned test data requirements.

[0115] In this embodiment, the electronic device can perform ALD determination for each test data point based on the Approximate Linear Dependence (ALD) algorithm, and obtain its approximate error value (ALD value). For example, the ALD value corresponding to each test data point can be calculated using the following formula:

[0116]

[0117] Where, x i (i = 1, K, z) represents the training data, z represents the number of training data, and x represents the training data. z+1 ε represents newly entered samples (test data) into the system. z+1 This means that the newly entered samples (test data) relative to the completed statistical limit T θ The error value between the established training data.

[0118] As can be seen, the ALD value ε corresponding to each test data point z+1 It can indicate the error value between the test data and the training data, so that the ALD value corresponding to the test data can be used to filter out the first test sub-data for model update and the second test sub-data that does not need to be used for model update.

[0119] As an optional implementation, an error threshold can be preset to define the degree of difference between the test data and the training data that can be used for model updates. For example, the operation of filtering the first and second test sub-data from the test data based on the ALD value can be represented by the following formula:

[0120]

[0121] Where v is the set error threshold.

[0122] In other words, when the ALD value ε of a certain test data is z+1 If the value is greater than the set threshold v, it indicates that this test data is relative to the statistical limit T that has been completed. θ The training data established are relatively independent, therefore statistical limits have been established. θ The training data cannot be linearly represented by the test data, therefore the test data needs to be fed into the training model to update the statistical limits and complete the real-time update of the model; when the ALD value ε of a certain test data is... z+1 If the value is not greater than the set threshold v, it indicates that the test data is relative to the statistical limit T. θ The established training data exhibit a certain linear relationship, which can be utilized by utilizing the pre-established statistical limits. θ The sample data is linearly represented by the test data, so there is no need to use the test data to update the model.

[0123] It should be noted that, in the embodiments of this application, the execution order of steps 150 and 160 is not logically necessarily related.

[0124] 170. Update the first autoregressive model using the first test sub-data to update the T statistical limit.

[0125] In this embodiment of the application, updating the first autoregressive model using the first test sub-data may include using the statistical limit T shown in equation (13). θ The first test data is included in the calculation.

[0126] 180. Using the KDLV algorithm, construct a second autoregressive model using the dynamic components in the second test sub-data, and calculate the T-statistic corresponding to the second autoregressive model.

[0127] In this embodiment, the specific implementation of step 180 can refer to the aforementioned step 150, including: the method of extracting the dynamic components of the second test sub-data can refer to the aforementioned step 1510; the method of building a second autoregressive model using the extracted dynamic components can refer to the aforementioned step 1520; and the method of calculating the T-statistic corresponding to the second autoregressive model can refer to the aforementioned step 1530. The following details will not be repeated. It is understood that the second autoregressive model can be any type of autoregressive model, and the T-statistic is based on T... 2 The statistical limit T of the data in the second autoregressive model obtained by statistical methods 2 .

[0128] 190. Determine the detection results of ice-related disasters on transmission lines based on the comparison results between the T-statistic limits corresponding to the training data and the T-statistic values ​​corresponding to the test data.

[0129] In this embodiment of the application, the statistical limits T of each data in the second autoregressive model are... 2 The statistical limit T calculated by the first autoregressive model θ By comparing the values, the freezing status of the transmission line at the time of detection can be determined. Specifically, if the T statistical value is less than or equal to the T statistical limit, it is determined that the transmission line is not subject to freezing. If the T statistical value is greater than the T statistical limit, it is determined that the transmission line is subject to freezing. Optionally, when freezing is determined to be present, the electronic equipment can output an alarm to remind inspection personnel.

[0130] In some possible embodiments, the aforementioned first and second autoregressive models are both AR(2) models. Since the establishment of the AR(2) model requires the dynamic latent variable values ​​at the current time and the two time points preceding it, problems arise during the fusion of the KDLV and ALD algorithms. For example, if a test data point is found to meet the update strategy after being judged by ALD rules, it is placed in the first autoregressive model to update the statistical limits. However, if a test data point following this test data point is found to not meet the update strategy after being judged by ALD rules, it is further decomposed into KDLV in the second autoregressive model. However, during the decomposition process, it is found that the subsequent test data point cannot be used to build the AR(2) model because the test data point preceding this test data point is in the first autoregressive model.

[0131] Therefore, when filtering the first and second test sub-data, the electronic device can further utilize the following rule: there must be at least three second test sub-data intervals between any two adjacent first test sub-data. That is, if a test data that meets the update condition shown in equation (15) does not have at least three second test sub-data intervals between it and the previous first test sub-data that do not meet the update condition, then the test data that meets the update condition shown in equation (15) will not be determined as the first test sub-data.

[0132] For example, the above rule can be represented as follows:

[0133] GZ = {gz} c |c=1,2,...H} (16)

[0134] GZ′=gz′ c |c=1,2,...H-1} (17)

[0135]

[0136] Where c represents the index of the test data that satisfies the update strategy, H represents the total number of test data that satisfies the update strategy, and gz` c This represents the difference between any two test data points that satisfy the update strategy. In this embodiment, points required by the GZY set are saved, while points required by the GZN set are considered invalid.

[0137] Thus, even if the first and second autoregressive models adopt the AR(2) model, the KDLV algorithm can be effectively combined with the ALD algorithm. Test data that do not meet the ALD rules can be effectively decomposed by the KDLV algorithm in the second autoregressive model to obtain its dynamic latent variable values ​​(dynamic components).

[0138] It is evident that the transmission line icing disaster detection method disclosed in this application can integrate the KDLV algorithm and the ALD model, and adaptively complete the early warning of extreme icing disasters of power grid transmission lines under the premise of using a small amount of icing weight data as a training set, thereby improving the accuracy of icing disaster detection.

[0139] Furthermore, in the data preprocessing, considering the temporal and seasonal specificities of the "solar intensity" characteristic, average solar intensity is used as meteorological data. Based on this, this embodiment can also filter temperature and humidity data by incorporating icing mechanisms. Therefore, this embodiment combines icing mechanisms with machine learning to detect icing disasters on transmission lines, effectively avoiding the limitations of using only mechanistic models for detection, and solving the problem that using only machine learning does not fully consider the influence of natural climatic factors in the formation of icing disasters.

[0140] Furthermore, considering the potential noise issues in the data, this application embodiment incorporates the VMD algorithm for data denoising. This enables the system to not only effectively filter noise during the data preprocessing stage but also to achieve adaptive detection while fully extracting the dynamic and static features of the data. Based on effective denoising, the input model exhibits high data validity and completeness, ensuring accurate detection of freezing disasters even with only a small amount of icing weight data as the training set. Moreover, it can detect extreme freezing disasters occurring once in a century, alleviating the computational difficulties and high complexity of previous methods.

[0141] Furthermore, during the fusion of the KDLV algorithm and the ALD model, there may be issues with the representation of the AR(2) model. This application's embodiment addresses this by setting conditional rules for screening test sub-data, enabling the KDLV algorithm to be effectively combined with the ALD algorithm. This allows for the full extraction of dynamic and static features of the data while simultaneously achieving adaptive detection of extreme freezing disasters on transmission lines, thus ensuring the practicality of the method.

[0142] Please see Figure 2 , Figure 2 This application discloses an embodiment of a transmission line icing disaster detection device. This device can be applied to any of the aforementioned electronic devices. Figure 2 As shown, the device includes:

[0143] The acquisition module 210 is used to acquire the first dataset; the first dataset includes icing weight data of transmission lines and meteorological condition data;

[0144] Preprocessing module 220 is used to preprocess the first dataset to obtain the second dataset;

[0145] The first processing module 230 is used to project the second dataset using the kernel principal component extraction strategy of the kernel dynamic latent variable (KDLV) algorithm to obtain the principal component matrix;

[0146] Extraction module 240 is used to extract training data and test data from the principal component matrix; wherein, the training data is the data in the principal component matrix representing the transmission line in normal state and in the first icing state; the test data is the data in the principal component matrix representing the transmission line in normal state and in the second icing state; the second icing state indicates a higher degree of icing than the first icing state;

[0147] The second processing module 250 is used to build a first autoregressive model using the dynamic components in the training data through the KDLV algorithm, and to calculate the T-statistic limit corresponding to the first autoregressive model.

[0148] The calculation module 260 is used to calculate the ALD value of each data in the test data, and to filter out the first test sub-data and the second test sub-data from the test data based on the ALD value;

[0149] Update module 270 is used to update the first autoregressive model using the first test sub-data in order to update the T statistical limit;

[0150] The third processing module 280 is used to build a second autoregressive model using the dynamic components in the second test sub-data through the KDLV algorithm, and to calculate the T-statistic corresponding to the second autoregressive model.

[0151] The determination module 290 is used to determine the detection results of ice disasters on transmission lines based on the comparison results of T statistical limits and T statistical values.

[0152] As an optional implementation, the determining module 290 is also used to output an alarm when the T statistical value is greater than the T statistical limit; the alarm is used to indicate that there is an ice disaster in the transmission line; or, when the T statistical value is less than or equal to the T statistical limit, it is determined that there is no ice disaster in the transmission line.

[0153] As an optional implementation, the calculation module 260 is further configured to filter out a first test sub-data and a second test sub-data from the test data based on the comparison result between the ALD value and the error threshold; wherein, the first test sub-data is the test data whose corresponding ALD value is greater than the error threshold; and the second test sub-data is the test data whose corresponding ALD value is greater than or equal to the error threshold.

[0154] As an optional implementation, both the first autoregressive model and the second autoregressive model are AR(2) models; there is at least a time interval of three second test data points between every two adjacent first test data points.

[0155] As an optional implementation, the preprocessing module 220 is further configured to decompose the data in the first dataset into K intrinsic mode functions using the variational mode decomposition (VMD) algorithm; where K is a positive integer greater than or equal to 1; and to extract the time-domain features and frequency features of the K intrinsic mode functions, and to standardize the extracted time-domain features and frequency-domain features respectively to obtain the second dataset.

[0156] As an optional implementation, the meteorological condition data includes at least the average sunshine intensity, which is the average of the sunshine intensity values ​​at N sampling times; N is a positive integer greater than or equal to 2. Further optionally, the meteorological condition data also includes temperature data within the range of -15°C to 0°C; or, the meteorological condition data also includes humidity data greater than 80%.

[0157] The transmission line icing detection device disclosed in the embodiments of this application can integrate the KDLV algorithm and the ALD model to adaptively complete the early warning of extreme icing disasters of power grid transmission lines under the premise of using a small amount of icing weight data as a training set, thereby improving the accuracy of icing disaster detection.

[0158] Please see Figure 3 , Figure 3 This is a schematic diagram of the structure of an electronic device disclosed in an embodiment of this application.

[0159] like Figure 3 As shown, the testing equipment may include:

[0160] Memory 310 storing computer programs;

[0161] Processor 320 coupled to memory 310;

[0162] When the computer program stored in the memory 310 is executed by the processor 320, the processor 320 implements any of the methods for detecting freezing of transmission lines disclosed in the embodiments of this application.

[0163] It should be noted that, Figure 3 The electronic device shown may also include components not shown, such as a power supply, input buttons, screen, RF circuit, Wi-Fi module, and Bluetooth module, which will not be described in detail in this embodiment.

[0164] This application discloses a computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements any of the methods for detecting freezing of transmission lines disclosed in this application.

[0165] This application discloses a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to execute any of the power transmission line freezing detection methods disclosed in this application.

[0166] It should be understood that the phrase "one embodiment" or "an embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of this application. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. Those skilled in the art should also recognize that the embodiments described in the specification are optional embodiments, and the actions and modules involved are not necessarily essential to this application.

[0167] In the various embodiments of this application, it should be understood that the sequence number of each process does not necessarily imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0168] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they can be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0169] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0170] If the aforementioned integrated units are implemented as software functional units and sold or used as independent products, they can be stored in a computer-accessible memory. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several requests to cause a computer device (which can be a personal computer, server, or network device, specifically a processor in the computer device) to execute some or all of the steps of the methods described in the various embodiments of this application.

[0171] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, including read-only memory (ROM), random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electrically-erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, disk storage, magnetic tape storage, or any other computer-readable medium capable of carrying or storing data.

[0172] The foregoing has provided a detailed description of the testing methods, testing equipment, testing systems, and storage media for telemetry devices disclosed in the embodiments of this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are merely for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for detecting freezing disasters on power transmission lines, characterized in that, The method includes: Obtain the first dataset; the first dataset includes icing weight data of transmission lines and meteorological condition data; The first dataset is preprocessed to obtain the second dataset; The second dataset is projected using the kernel principal component extraction strategy of the kernel dynamic latent variable (KDLV) algorithm to obtain the principal component matrix; Training data and test data are extracted from the principal component matrix; wherein, the training data is the data in the principal component matrix representing the transmission line in a normal state and in a first icing state; the test data is the data in the principal component matrix representing the transmission line in a normal state and in a second icing state; the second icing state indicates a higher degree of icing than the first icing state; The first autoregressive model is constructed using the dynamic components in the training data through the KDLV algorithm, and the T-statistic limit corresponding to the first autoregressive model is calculated. Calculate the ALD value of each data point in the test data, and filter out the first test sub-data and the second test sub-data from the test data based on the ALD value; The first autoregressive model is updated using the first test sub-data to update the T statistical limit; Using the KDLV algorithm, a second autoregressive model is constructed using the dynamic components in the second test sub-data, and the T-statistic corresponding to the second autoregressive model is calculated. The detection results of the transmission line freezing disaster are determined based on the comparison between the T statistical limit and the T statistical value; The step of filtering the first test sub-data and the second test sub-data from the test data based on the ALD value includes: Based on the comparison results between the ALD value and the error threshold, the first test sub-data and the second test sub-data are selected from the test data; The first test sub-data is the test data whose corresponding ALD value is greater than the error threshold; the second test sub-data is the test data whose corresponding ALD value is less than or equal to the error threshold. Both the first autoregressive model and the second autoregressive model are AR(2) models; there is at least a time interval of three second test data points between every two adjacent first test data points.

2. The method according to claim 1, characterized in that, The determination of the detection results of transmission line freezing disaster based on the comparison result of the T statistical limit and the T statistical value includes: If the T statistical value is greater than the T statistical limit, an alarm is output; the alarm is used to indicate that there is freezing damage to the transmission line; or, If the T statistical value is less than or equal to the T statistical limit, then it is determined that there is no freezing disaster on the transmission line.

3. The method according to any one of claims 1-2, characterized in that, The preprocessing of the first dataset to obtain the second dataset includes: The data in the first dataset is decomposed into K eigenmode functions using the Variational Mode Decomposition (VMD) algorithm; K is a positive integer greater than or equal to 1. The time-domain and frequency-domain features of the K intrinsic mode functions are extracted, and the extracted time-domain and frequency-domain features are standardized to obtain the second dataset.

4. The method according to any one of claims 1-2, characterized in that, The meteorological condition data includes at least the average sunshine intensity, which is the average of the sunshine intensity values ​​at N sampling times; N is a positive integer greater than or equal to 2.

5. The method according to claim 4, characterized in that, The meteorological conditions data shall include at least temperature data in the range of -15°C to 0°C; or, the meteorological conditions data shall include at least humidity data greater than 80%.

6. A device for detecting freezing disasters on power transmission lines, characterized in that, For implementing the method as described in any one of claims 1-5, comprising: The acquisition module is used to acquire a first dataset; the first dataset includes icing weight data of transmission lines and meteorological condition data; The preprocessing module is used to preprocess the first dataset to obtain the second dataset; The first processing module is used to project the second dataset using the kernel principal component extraction strategy of the kernel dynamic latent variable (KDLV) algorithm to obtain the principal component matrix. An extraction module is used to extract training data and test data from the principal component matrix; wherein, the training data is data in the principal component matrix representing the transmission line in a normal state and in a first icing state; the test data is data in the principal component matrix representing the transmission line in a normal state and in a second icing state; the second icing state indicates a higher degree of icing than the first icing state; The second processing module is used to build a first autoregressive model using the dynamic components in the training data through the KDLV algorithm, and to calculate the T-statistic limit corresponding to the first autoregressive model. The calculation module is used to calculate the ALD value of each data in the test data, and to filter out the first test sub-data and the second test sub-data from the test data based on the ALD value; The update module is used to update the first autoregressive model using the first test sub-data in order to update the T statistical limit; The third processing module is used to build a second autoregressive model using the dynamic components in the second test sub-data through the KDLV algorithm, and to calculate the T-statistic corresponding to the second autoregressive model. The determination module is used to determine the detection results of ice disasters on transmission lines based on the comparison results of the T statistical limit and the T statistical value.

7. An electronic device, characterized in that, The electronic device includes a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor causes the processor to implement the method as described in any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 5.

Images