A method for identifying icing patterns of transmission lines

Abstract

The present invention discloses a method for identifying icing patterns on power transmission lines, comprising: generating an ice thickness sequence curve based on a historical icing data sequence of the transmission line; obtaining multiple icing event data based on the ice thickness sequence curve; and preprocessing each icing event data to obtain preprocessed icing event data, wherein the icing data includes ice thickness data and meteorological data, wherein the meteorological data includes at least temperature and relative humidity; extracting several characteristic parameters from the ice thickness sequence curve; calculating characteristic parameter values ​​of each preprocessed icing event data on each characteristic parameter to obtain corresponding characteristic vectors, and forming a characteristic matrix from the characteristic vectors; and performing hierarchical weighted clustering on the characteristic matrix according to a hierarchical weighted clustering method to identify the type of icing pattern on the transmission line. The present invention can accurately identify the characteristics of ice thickness variations within an icing event, facilitates the operation and maintenance of power transmission lines in icy and snowy weather, and provides technical support for anti-icing and anti-icing of power transmission lines.

Description

Technical Field

[0001] The present invention relates to the technical field of power transmission and transformation, and in particular to a method for identifying icing patterns of power transmission lines. Background Art

[0002] Icing on transmission lines can cause ice flashes, ice dancing, and even ground wire breakage and tower collapse, damaging transmission line infrastructure and threatening the safe and stable operation of the power system. An icing event refers to the entire process from ice accumulation to complete deicing on a transmission line. The icing pattern type reflects how ice thickness changes over time during this process. Accurately identifying the icing pattern type is crucial for power grid anti-icing efforts, line ice thickness prediction, and line anti-icing design.

[0003] With the development of transmission line icing monitoring technology, a large amount of icing monitoring data has been accumulated, making it possible to use big data technology to analyze the types of icing patterns on transmission lines. However, most current icing prediction methods only break down the icing process into different stages. They do not systematically describe the changes in ice thickness in each stage, or even in all stages, nor do they further identify the icing pattern formed by the entire icing event. Therefore, an effective method for identifying the type of icing patterns on transmission lines is urgently needed. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for identifying the icing pattern of a transmission line, so as to solve the technical problem in the prior art that the characteristics of ice thickness changes in each stage or even all stages of the icing process cannot be identified.

[0005] The purpose of the present invention can be achieved through the following technical solutions:

[0006] A method for identifying icing patterns of transmission lines, comprising:

[0007] generating an ice thickness sequence curve based on a historical ice data sequence of the transmission line, obtaining a plurality of ice event data based on the ice thickness sequence curve, and preprocessing each of the ice event data to obtain preprocessed ice event data, wherein the ice data in the ice data sequence includes ice thickness data and meteorological data, and the meteorological data includes at least temperature and relative humidity;

[0008] Extracting a number of characteristic parameters from the ice thickness sequence curve, calculating characteristic parameter values ​​of each pre-processed ice event data on each characteristic parameter to obtain corresponding characteristic vectors, and forming a characteristic matrix with the characteristic vectors;

[0009] The characteristic matrix is ​​subjected to hierarchical weighted clustering according to a hierarchical weighted clustering method to identify the type of icing law of the transmission line.

[0010] Optionally, determining a plurality of icing event data according to the icing thickness sequence curve includes:

[0011] When there is meteorological data at the same time as the ice thickness data, the ice thickness sequence curve between the start and end points is selected as an ice event data, with 3≤M, N≤4;

[0012] When there is no meteorological data at the same time as the ice thickness data, the ice thickness sequence curve between the start and end points is selected as an ice event data, with 3≤M, N≤4, taking the first M hours before the ice thickness starts to increase from 0 as the starting point and the last N hours after the ice is completely shed and the thickness reaches 0 as the end point.

[0013] Optionally, performing hierarchical weighted clustering on the feature matrix according to a hierarchical weighted clustering method includes:

[0014] First-layer clustering: using a hierarchical weighted clustering method to divide the pre-processed icing event data into single ice peak type and double ice peak type according to a first feature parameter set and a preset first set of weight coefficients;

[0015] Second-level clustering: using a hierarchical weighted clustering method to classify the single ice peak type into a single ice peak saturation type and a spike type according to a second characteristic parameter set and a preset second set of weight coefficients; and to classify the double ice peak type into a double ice peak artificial melting type and a double ice peak natural oscillation type according to a third characteristic parameter set and a preset third set of weight coefficients;

[0016] The third layer of clustering: using the hierarchical weighted clustering method according to the fourth feature parameter quantity The fourth group of weight coefficients is set and preset to divide the peak type into a single ice peak rapid ice accretion type, a single ice peak uniform ice accretion type and a single ice peak rapid ice shedding type;

[0017] Among them, the first feature parameter set, the second feature parameter set, the third feature parameter set and the fourth feature parameter set are subsets of the feature vector.

[0018] Optionally, the feature parameter set includes:

[0019] The time t corresponding to the peak ice thickness max , curve area A, time P corresponding to the second ice peak t , ice thickness P corresponding to the second ice peak d , Ice Valley corresponding time Q t , Ice thickness Q corresponding to the ice valley d , the time U corresponding to the maximum ice accretion rate t , ice thickness U corresponding to the maximum ice accumulation rate d , the time V corresponding to the maximum deicing rate t, ice thickness V corresponding to the maximum deicing rate d , the saturation time S during which the ice thickness maintains a saturated state, and the curve area A is the discrete trapezoidal integral area of ​​the closed area between the ice thickness sequence curve and the time axis.

[0020] Optionally, the first feature parameter set includes:

[0021] Each pre-processed ice event data is at the time P corresponding to the second ice peak t , ice thickness P corresponding to the second ice peak d , Ice Valley corresponding time Q t , Ice thickness Q corresponding to the ice valley d The characteristic parameter value on .

[0022] Optionally, the second feature parameter set includes:

[0023] The characteristic parameter values ​​of each preprocessed icing event data on the curve area A and the saturation time S during which the ice thickness maintains a saturated state.

[0024] Optionally, the third feature parameter set includes:

[0025] The ice thickness P corresponding to the second ice peak of each pre-processed icing event data d , Ice thickness Q corresponding to the ice valley d The characteristic parameter value on .

[0026] Optionally, the fourth feature parameter set includes:

[0027] The pre-processed ice event data at the time t corresponding to the peak ice thickness max , the time U corresponding to the maximum ice accretion rate t , ice thickness U corresponding to the maximum ice accumulation rate d , the time V corresponding to the maximum deicing rate t , ice thickness V corresponding to the maximum deicing rate d The characteristic parameter value on .

[0028] Optionally, preprocessing each of the ice event data includes:

[0029] The data of each ice event are normalized, outliers are eliminated, interpolated, and smoothed.

[0030] Optionally, the hierarchical weighted clustering method is a K-means weighted clustering method.

[0031] In view of this, the beneficial effects brought by the present invention are:

[0032] The present invention obtains multiple icing event data based on historical icing data of power transmission lines. Taking into account the potential noise in icing data due to factors such as breeze vibration and sensor accuracy, the icing event data is preprocessed to ensure the accuracy of icing patterns at the source. To address the challenges of a large number of icing event data sequences and varying sequence lengths, several characteristic parameters are extracted to characterize ice thickness sequence curves. The characteristic parameter values ​​for each characteristic parameter of the preprocessed icing event data are calculated to obtain corresponding characteristic vectors. Using a hierarchical weighted clustering method and characteristic vectors, the preprocessed icing event data is hierarchically weighted clustered to accurately identify icing pattern types, overcoming the challenges of an unclear number of icing pattern types and varying characterization parameters. The method provided by the present invention reveals the ice thickness variation process of an icing event from the onset of ice accumulation to the completion of deicing. It can accurately identify the characteristics of ice thickness variation within an icing event, facilitating the operation and maintenance of actual transmission lines in icy and snowy weather and providing technical support for anti-icing and anti-icing work on transmission lines. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 Schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0034] The embodiment of the present invention provides a method for identifying the icing pattern of a transmission line to solve the technical problem in the prior art that it is impossible to identify the characteristics of ice thickness changes in each stage or even all stages of the icing process.

[0035] To facilitate understanding of the present invention, the present invention will be described more fully below with reference to the accompanying drawings. Preferred embodiments of the present invention are shown in the drawings. However, the present invention may be embodied in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a more thorough and comprehensive disclosure of the present invention.

[0036] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this invention pertains. The terms used in this specification of the present invention are for the purpose of describing specific embodiments only and are not intended to limit the present invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0037] Most current methods only decompose the icing process into different stages, but do not further describe the characteristics of ice thickness changes in each stage or even all stages, and do not further summarize the types of icing rules formed by the entire icing event. To this end, the present invention provides a method for identifying icing rules for transmission lines based on the monitoring data of the transmission line icing warning system. In order to solve the problem of noise in the measurement data caused by breeze vibration, sensor accuracy, etc., the local weighted scatter point smoothing (LOWESS) method is used to smooth the curve; in order to solve the problem of unequal lengths of time series of icing events and a large number of series, 11 characteristic parameters are proposed to characterize the ice thickness time curve; in order to solve the problem of unclear number of types and unequal strength of characterization parameters, the K-means hierarchical weighted clustering method is proposed to cluster icing events to obtain icing rule types. The method provided by the present invention reveals the process of ice thickness changes from the beginning to the end of an icing event, which is closer to the actual operation and maintenance needs of transmission lines in ice and snow weather, and provides technical support for the anti-icing and anti-icing work of transmission lines.

[0038] See also Figure 1 The following is an embodiment of a method for identifying ice coating patterns on a transmission line according to the present invention, comprising:

[0039] S100: generating an ice thickness sequence curve based on a historical ice data sequence of a transmission line, obtaining a plurality of ice event data based on the ice thickness sequence curve, and preprocessing each of the ice event data to obtain preprocessed ice event data, wherein the ice data in the ice data sequence includes ice thickness data and meteorological data, wherein the meteorological data includes at least temperature and relative humidity;

[0040] S200: extracting a plurality of characteristic parameters from the ice thickness sequence curve, calculating characteristic parameter values ​​of each pre-processed ice event data on each characteristic parameter to obtain corresponding characteristic vectors, and forming a characteristic matrix with the characteristic vectors;

[0041] S300: performing hierarchical weighted clustering on the characteristic matrix according to a hierarchical weighted clustering method to identify the type of icing pattern of the transmission line.

[0042] In this embodiment, step S100 obtains multiple icing event data based on the icing data sequence of the transmission line over the years, and preprocesses the icing event data to obtain preprocessed icing event data; wherein the icing data in the icing data sequence includes ice thickness data and meteorological data, and the meteorological data includes at least temperature and relative humidity. Specifically, the process of obtaining multiple icing event data is as follows:

[0043] (1) Collecting icing data: Collect the ice thickness data series and corresponding meteorological data series of each line terminal in the overhead line icing online monitoring system over the years, where the meteorological data mainly includes temperature and relative humidity.

[0044] (2) Selecting ice event data: Plot ice data (e.g., ice thickness, temperature, and relative humidity) into an ice thickness sequence curve (referred to as an ice thickness sequence curve). Selecting ice event data is divided into two cases: with and without meteorological conditions. When there are temperature and relative humidity data at the same time as the ice thickness, the ice thickness sequence curve corresponding to t1 to t2 is selected as a complete ice event data, starting from t1 for the first M hours before the temperature drops to 0°C and ending at t2 for the N hours after the temperature rises to 0°C. When there are no temperature and relative humidity data at the same time as the ice thickness, the ice thickness sequence curve corresponding to t1 to t2 is selected as a complete ice event data, starting from t1 for the first M hours before the ice thickness starts to increase from 0 and ending at t2 for the N hours after the ice is completely shed and becomes 0. Wherein, 3≤M, N≤4, and in a preferred embodiment, M=N=3.

[0045] (3) Verification of icing event data: When all the following conditions are met, the icing event data is of high quality and is retained; otherwise, it is discarded: the maximum ice thickness max(b) is greater than or equal to 2 mm; the total duration of the icing event data is greater than or equal to 18 h; the time difference values ​​diff(t) before and after all icing data series are less than 24 h; the total number of icing data series is greater than or equal to 30; the proportion of data null points in the total number is less than 10%.

[0046] Specifically, the process of preprocessing icing event data is as follows:

[0047] (1) Normalization: The ice event data are mapped to the interval [0, 1]. After normalization, only the middle main part and the 0 values ​​closest to the main part are retained for the ice thickness series curve, and other redundant 0 values ​​are discarded.

[0048] (2) Outlier elimination and interpolation processing: Use box plot to filter outliers in the sequence, and use polynomial interpolation method to improve the data for outliers and missing values.

[0049] (3) Curve smoothing: Local weighted scatter point smoothing (LOWESS) is used to smoothly fit the curve corresponding to the icing event data, which significantly eliminates the noise of the curve.

[0050] In step S200, several characteristic parameters are extracted from the ice thickness sequence curve, and the characteristic parameter values ​​of each pre-processed ice event data on each characteristic parameter are calculated to obtain the corresponding characteristic vector. In this embodiment, the ice thickness sequence curve is analyzed, marked according to the analyzed characteristic parameters, and several characteristic parameters are extracted. In a preferred embodiment, 11 characteristic parameters of the ice thickness sequence curve are extracted, namely: the time t corresponding to the ice thickness peak max , curve area A, the time P corresponding to the second ice peak t , ice thickness P corresponding to the second ice peak d , Ice Valley corresponding time Q t , Ice thickness Q corresponding to the ice valley d , the time U corresponding to the maximum ice accretion rate t , ice thickness U corresponding to the maximum ice accumulation rate d , the time V corresponding to the maximum deicing rate t , ice thickness V corresponding to the maximum deicing rate d , ice thickness maintains saturation state saturation time S. It is worth noting that the second ice peak P is expressed by (P t ,P d ) indicates that Ice Valley Q is represented by (Q t ,Q d ), the maximum ice accretion rate U is expressed as (U t ,U d ) is expressed as, the maximum deicing rate V is expressed as (V t ,V d )express.

[0051] Specifically, the textual descriptions and mathematical definitions of the 11 characteristic parameters of the ice thickness series curve are as follows:

[0052] (1)t max is the time corresponding to the peak ice thickness.

[0053]

[0054] Among them, t i Indicates the ice thickness data d i The corresponding time, 1≤i≤n, n represents the number of ice thickness data series, t max The range of is [0,1].

[0055] (2) Curve area A: It is the discrete trapezoidal integral area of ​​the closed area between the ice thickness series curve and the time axis.

[0056]

[0057] The range of the curve area A is [0,1].

[0058] (3) The time P corresponding to the second ice peak t , ice thickness P corresponding to the second ice peak d , Ice Valley corresponding time Q t , Ice thickness Q corresponding to the ice valley d In a given ice thickness data sequence, if there is a local maximum d i (excluding the global maximum (t max ,d max )) and a local minimum d j , and d i –d j >th1, where th1 is the adjacent peak-to-valley difference threshold, then:

[0059]

[0060] If the two local extreme points mentioned above do not exist, then:

[0061]

[0062] The value range of all parameters of ice peaks and ice valleys is [0,1].

[0063] (4) The time U corresponding to the maximum ice accretion rate t , ice thickness U corresponding to the maximum ice accumulation rate d , the time V corresponding to the maximum deicing rate t , ice thickness V corresponding to the maximum deicing rate d It is worth noting that the maximum ice accretion rate U is expressed as (U t ,U d ) is expressed as, the maximum deicing rate V is expressed as (V t ,V d ). After normalization, the time length is 1. The curve is divided according to the step size th2. The mean ice thickness at equal steps is calculated. The maximum and minimum values ​​of the first-order difference of the mean series are taken as the maximum ice accretion rate and the maximum ice shedding rate, respectively.

[0064]

[0065] The range of the time in the maximum ice accretion rate U is [0,1]. Similarly, the range of the time in the maximum ice shedding rate V is [0,1].

[0066] (5) Saturation duration S: refers to the time duration that the ice thickness maintains saturation near the peak value. m ,t n ], satisfying t max ∈[t m ,t n ] and max{d m ,d m+1,…,d n}–min{d m ,d m+1 ,…,d n}<th3, where th3 is the maximum interval ice thickness difference threshold, then:

[0067] S=t n -t m (6)

[0068] The value range of the saturation duration S is [0,1].

[0069] For each icing event, the characteristic parameter values ​​for the 11 characteristic parameters of the ice thickness series curve are calculated. All characteristic parameters are combined into a characteristic vector, as shown in formula (7). This characteristic vector can represent the curve corresponding to the icing event data. It can be understood that each icing event data has a corresponding characteristic vector, which includes 11 characteristic parameter values.

[0070] r=(t max AP t P d Q t Q d U t U d V t V d S) (7)

[0071] When there are n ice event data, there are corresponding n eigenvectors. These n eigenvectors can form a feature matrix, see formula (8), and the feature matrix can be conveniently used in clustering algorithms.

[0072]

[0073] In step S300, hierarchical weighted clustering is performed on the feature matrix extracted from the pre-processed icing event data according to a hierarchical weighted clustering method to identify the icing pattern type of the transmission line.

[0074] Based on the feature matrix extracted from the pre-processed icing event data, a hierarchical weighted clustering method is used. In the preferred embodiment, the weighted clustering method is the K-means hierarchical weighted clustering method. The icing event data are subjected to three-layer weighted K-means clustering, and the icing pattern types are divided into 6 types, namely: single ice peak saturation type (type I), single ice peak rapid icing type (type II), single ice peak uniform icing type (type III), single ice peak rapid icing type (type IV), double ice peak artificial ice melting type (type V), and double ice peak natural oscillation type (type VI).

[0075] (1) First-level clustering: The pre-processed icing event data is divided into single ice peak type and double ice peak type according to the first characteristic parameter set and the preset first set of weight coefficients using a hierarchical weighted clustering method. The first-level clustering aims to divide the pre-processed icing event data into single ice peak curve and double ice peak curve. The monotonic ice accumulation and de-icing process will form an ice thickness sequence curve of a single ice peak, while the staggered ice accumulation and de-icing process will form an ice thickness sequence curve of a double ice peak and a single ice valley. The difference in parameters between the two is mainly reflected in the two parameters of ice peak P and ice valley Q. For the single ice peak curve, P and Q are 0, while for the double ice peak and single ice valley curve, P and Q are not 0.

[0076] Therefore, P is selected for the first-level clustering. t 、P d , Q t , Q d As characteristic parameters, the first characteristic parameter set includes: each pre-processed ice event data at the time P corresponding to the second ice peak t , ice thickness P corresponding to the second ice peak d , Ice Valley corresponding time Q t , Ice thickness Q corresponding to the ice valley d The first set of weight coefficients is shown in formula (9). The characteristic parameter values ​​in the matrix R1 are multiplied by the first set of weight coefficients in coeff1 to form an n×4 matrix consisting only of P and Q for K-means weighted clustering.

[0077]

[0078] According to formula (9), we can know that P t The corresponding weight coefficient is C Pt , P d The corresponding weight coefficient is C Pd , Q t The corresponding weight coefficient is C Qt , Q d The corresponding weight coefficient is C Qd ; In a preferred embodiment, the first set of weight coefficients is (0.3, 0.3, 0.2, 0.2).

[0079] (2) Second-level clustering: Using a hierarchical weighted clustering method, the single ice peak type is divided into a single ice peak saturation type and a spike type according to the second characteristic parameter set and a preset second set of weight coefficients. The double ice peak type is divided into a double ice peak artificial melting type and a double ice peak natural oscillation type according to the third characteristic parameter set and a preset third set of weight coefficients. The second-level clustering aims to classify the saturation curve and the spike curve in the single ice peak curve, and the artificial melting curve and the natural oscillation curve in the double ice peak curve, and to classify the single ice peak curve into the single ice peak saturation type and the spike type, and the double ice peak curve into the double ice peak artificial melting type and the double ice peak natural oscillation type.

[0080] For the single ice peak curve, the length of time that saturation is maintained near the peak will greatly affect the shape of the ice thickness sequence curve. A long saturation time will flatten the ice thickness sequence curve, while a short saturation time will make the ice thickness sequence curve form a spike. The difference in parameters between the two is mainly reflected in the saturation time S. At the same time, under normalized conditions, a long saturation time makes the area under the curve A larger, while the peak curve is relatively smaller. Therefore, when the second-level clustering is performed on the single ice peak type, A and S are selected as feature parameters. The second feature parameter set includes: the feature parameter values ​​of each pre-processed icing event data on the curve area A and the saturation time S for ice thickness to maintain a saturated state. The second set of weight coefficients is shown in formula (10). The matrix R 2-A The characteristic parameter value in is multiplied by coeff 2-A The weight coefficients in form an n×2 matrix consisting only of A and S for K-means weighted clustering.

[0081]

[0082] According to formula (10), the weight coefficient corresponding to A is C A , the weight coefficient corresponding to S is C S ; In a preferred embodiment, the second set of weight coefficients is (0.2, 0.8).

[0083] For the double ice peak curve, although both have the process of ice accumulation, de-icing and ice accumulation, there are generally two types of de-icing during the ice accumulation process in the natural environment. One is large-scale de-icing caused by artificial ice melting. In this case, the ice thickness sequence curve will show a sharp drop or even drop to 0; the other is small-scale de-icing caused by short-term warming, strong winds or mechanical vibrations. In this case, the ice thickness sequence curve generally only shows a slight drop and forms an oscillating curve. The difference in parameters between the above two cases is mainly reflected in the ice thickness P corresponding to the second ice peak. d Ice thickness Q corresponding to the ice valley d Understandably, P d is the amplitude of the second ice peak, Q d is the amplitude of the ice valley. Therefore, when the second-level clustering is for double ice peak clustering, P is selected. d , Q d As characteristic parameters, the third characteristic parameter set includes: the ice thickness P corresponding to the second ice peak of each pre-processed icing event data d , Ice thickness Q corresponding to the ice valley d The characteristic parameter value on the third group of weights is shown in formula (11), and the matrix R 2-B The characteristic parameter value in is multiplied by coeff 2-B The weight coefficient in is formed by only P d , Q dThe n×2 matrix composed of the K-means weighted clustering is performed.

[0084]

[0085] According to formula (11), we can know that P d The corresponding weight coefficient is C Pd , Q d The corresponding weight coefficient is C Qd ; In a preferred embodiment, the third set of weight coefficients is (-0.2, 1.2).

[0086] The third level of clustering: using the hierarchical weighted clustering method, the peak type is divided into a single ice peak rapid ice accumulation type, a single ice peak uniform ice accumulation type and a single ice peak rapid ice shedding type according to the fourth characteristic parameter set and the preset fourth set of weight coefficients. The third level of clustering aims to refine the peak formation period of the peak curve in the single ice peak type. For the single ice peak peak curve, although they all have peak characteristics, the peak formation period can be further divided into the early, middle and late stages. If the peak is formed in the early stage, it means that the transmission line has experienced a process of rapid ice accumulation followed by slow ice shedding, which is a single ice peak rapid ice accumulation type; similarly, if the ice is formed in the late stage, it means that the transmission line first slowly accumulates ice and then quickly sheds ice, which is a single ice peak rapid ice shedding type; if the ice is formed in the middle stage, the rates of ice accumulation and ice shedding may be the same, which is a single ice peak uniform ice accumulation type. The difference in parameters between the above three cases is mainly reflected in the time t corresponding to the peak value of the ice thickness. max , maximum ice accretion rate U and maximum ice shedding rate V.

[0087] Therefore, the third-level clustering selects t max 、U t 、U d 、V t 、V d As characteristic parameters, the fourth characteristic parameter set includes: the time t corresponding to the peak value of ice thickness of each pre-processed ice event data max , the time U corresponding to the maximum ice accretion rate t , ice thickness U corresponding to the maximum ice accumulation rate d , the time V corresponding to the maximum deicing rate t , ice thickness V corresponding to the maximum deicing rate d The fourth set of weight coefficients is shown in formula (12), which multiplies the characteristic parameter values ​​in the matrix R3 by the weight coefficients in coeff3 to form a matrix consisting only of t max 、U t 、U d 、V t 、V d The n×5 matrix composed of the K-means weighted clustering is performed.

[0088]

[0089] According to formula (12), we can know that t max The corresponding weight coefficient is C tmax , U t The corresponding weight coefficient is C Ut , U d The corresponding weight coefficient is C Ud , V t The corresponding weight coefficient is C Vt , V d The corresponding weight coefficient is C Vd ; In a preferred embodiment, the fourth group of weight coefficients is (0.6, 0.1, 0.1, 0.1, 0.1).

[0090] It can be understood that the first feature parameter set, the second feature parameter set, the third feature parameter set and the fourth feature parameter set are subsets of the feature vector.

[0091] The method for identifying icing patterns on power transmission lines provided in this embodiment obtains multiple icing event data based on historical icing data of transmission lines. Taking into account the potential for noise in the icing data due to factors such as breeze vibration and sensor accuracy, the icing event data is preprocessed to ensure the accuracy of the icing patterns at the source. To address the challenges of a large number of icing event data sequences and varying sequence lengths, several characteristic parameters are extracted to characterize the ice thickness sequence curve. The characteristic parameter values ​​for each characteristic parameter of the preprocessed icing event data are calculated to obtain corresponding characteristic vectors. Using a hierarchical weighted clustering method and characteristic vectors, the preprocessed icing event data is subjected to hierarchical weighted clustering, accurately identifying the type of icing pattern, overcoming the challenges of an unclear number of icing pattern types and varying characterization parameters. The method disclosed herein reveals the ice thickness variation process throughout an icing event, from the onset of ice accumulation to the completion of deicing. It can accurately identify the characteristics of ice thickness variation within an icing event, facilitating the operation and maintenance of actual transmission lines in icy and snowy weather and providing technical support for anti-icing and anti-icing efforts on transmission lines.

[0092] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0093] In the embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0094] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0095] In addition, the functional units in the various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0096] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, 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. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0097] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for identifying ice coverage patterns of power transmission lines, characterized in that: include: generating an ice thickness sequence curve based on a historical ice data sequence of the transmission line, obtaining a plurality of ice event data based on the ice thickness sequence curve, and preprocessing each of the ice event data to obtain preprocessed ice event data, wherein the ice data in the ice data sequence includes ice thickness data and meteorological data, and the meteorological data includes at least temperature and relative humidity; Extracting a number of characteristic parameters from the ice thickness sequence curve, calculating characteristic parameter values ​​of each pre-processed ice event data on each characteristic parameter to obtain corresponding characteristic vectors, and forming a characteristic matrix with the characteristic vectors; Performing hierarchical weighted clustering on the characteristic matrix according to a hierarchical weighted clustering method to identify the type of icing law of the transmission line; The step of performing hierarchical weighted clustering on the feature matrix according to the hierarchical weighted clustering method comprises: First-layer clustering: using a hierarchical weighted clustering method to divide the pre-processed icing event data into single ice peak type and double ice peak type according to a first feature parameter set and a preset first set of weight coefficients; Second-level clustering: using a hierarchical weighted clustering method to classify the single ice peak type into a single ice peak saturation type and a spike type according to a second characteristic parameter set and a preset second set of weight coefficients; and to classify the double ice peak type into a double ice peak artificial melting type and a double ice peak natural oscillation type according to a third characteristic parameter set and a preset third set of weight coefficients; Third-level clustering: using a hierarchical weighted clustering method according to a fourth characteristic parameter set and a preset fourth set of weight coefficients, the peak type is divided into a single ice peak rapid ice accretion type, a single ice peak uniform ice accretion type, and a single ice peak rapid ice shedding type; Among them, the first characteristic parameter set, the second characteristic parameter set, the third characteristic parameter set and the fourth characteristic parameter set are subsets of the characteristic vector; the first characteristic parameter set includes: each pre-processed icing event data at the time P corresponding to the second ice peak t , ice thickness P corresponding to the second ice peak d , Ice Valley corresponding time Q t , Ice thickness Q corresponding to the ice valley d The second characteristic parameter set includes: the characteristic parameter values ​​of each pre-processed icing event data on the curve area A, the saturation time S of the ice thickness maintaining the saturation state; the third characteristic parameter set includes: the ice thickness P corresponding to the second ice peak of each pre-processed icing event data d , Ice thickness Q corresponding to the ice valley d The fourth characteristic parameter set includes: each pre-processed icing event data at the time t corresponding to the peak value of the icing thickness max , the time U corresponding to the maximum ice accumulation rate t , ice thickness U corresponding to the maximum ice accumulation rate d , the time V corresponding to the maximum deicing rate t , ice thickness V corresponding to the maximum deicing rate d The characteristic parameter value on the curve; the curve area A is the discrete trapezoidal integral area of ​​the ice thickness sequence curve and the time axis closed area.

2. The method for identifying ice coverage patterns of power transmission lines according to claim 1, characterized in that: Determining a plurality of icing event data according to the icing thickness sequence curve includes: When there is meteorological data at the same time as the ice thickness data, the ice thickness sequence curve between the start and end points is selected as an ice event data, with 3≤M, N≤4; When there is no meteorological data at the same time as the ice thickness data, the ice thickness sequence curve between the start and end points is selected as an ice event data, with 3≤M, N≤4, taking the first M hours before the ice thickness starts to increase from 0 as the starting point and the last N hours after the ice is completely shed and the thickness reaches 0 as the end point.

3. The method for identifying ice coverage patterns of power transmission lines according to claim 1, characterized in that: Preprocessing the ice event data includes: The data of each ice event are normalized, outliers are eliminated, interpolated, and smoothed.

4. The method for identifying icing patterns of power transmission lines according to any one of claims 1 to 3, characterized in that: The hierarchical weighted clustering method is a K-means hierarchical weighted clustering method.

Images