Tower load early warning method and device, terminal and storage medium

Abstract

The present application relates to the technical field of tower protection, and particularly relates to a tower load early warning method, device, terminal and storage medium, the method of the present application first acquires a plurality of factor data sets; then the plurality of factor data sets are subjected to feature extraction and weighting processing to obtain a factor vector; then the class center with the minimum distance from the factor vector is determined as the target class center; finally, according to the target class center, the class radius of the influence class and the distance between the factor vector and the target class center, the load early warning level is determined. The present application determines a feature filtering vector based on the factor data set using a correlation analysis method, extracts useful features based on the filtering vector, constructs a linear load model based on the features, determines the coefficients of each feature through the load model, constructs a factor vector, and finally determines the early warning level according to the belonging class through the classification method. Therefore, the early warning level division is more scientific and has more basis, thereby solving the problem that the risk amount caused by external factors is not easy to be determined.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of tower protection, and in particular to a tower load early warning method and device, a terminal and a storage medium. BACKGROUND

[0002] A tower is a pole-shaped or tower-shaped structure that supports overhead transmission lines and overhead ground wires and maintains a certain distance between them and the ground. Steel structures, wood structures and reinforced concrete structures are mostly used for line towers in the world. In general, the pole-shaped structure of wood and reinforced concrete is called a pole, and the tower-shaped steel structure and reinforced concrete chimney-shaped structure is called a tower. A tower without a guy wire is called a self-supporting tower, and a tower with a guy wire is called a guyed tower.

[0003] A tower structure is a statically indeterminate structure, and the failure of a certain member does not cause the failure of the entire structure. Only when the number of damaged members reaches a certain number, the tower can no longer bear the load, and the tower is considered to be damaged. Due to the existence of a large number of uncertain factors in the tower structure, for example, the load acting on the tower is mainly wind load, ice and snow load, earthquake load and conductor self-weight load, which are all dynamic loads varying with time.

[0004] In fact, the health status of the tower is closely related to external factors, for example, extreme weather such as gales, severe cold and snow can cause damage to the tower, and in severe cases can even cause the tower to be damaged. If the analysis and early warning of the tower are made in time according to the external factors combined with the tower itself, appropriate maintenance measures can be taken, which is of great significance for the protection of the tower.

[0005] Therefore, it is necessary to develop a tower load early warning method. SUMMARY

[0006] The embodiments of the present application provide a tower load early warning method, device, terminal and storage medium, which are used to solve the problem that the risk amount caused by external factors to the tower is not easy to be determined in the prior art.

[0007] In a first aspect, the embodiments of the present application provide a tower load early warning method, comprising:

[0008] Obtaining a plurality of factor data sets, wherein the plurality of factor data sets are obtained based on the same time period, and each factor data set includes a plurality of factor data corresponding to a plurality of time nodes, and the factor data affects the tower body load;

[0009] Performing feature extraction and weighting processing on the plurality of factor data sets to obtain a factor vector;

[0010] determining a class center with the minimum distance from the factor vector as a target class center from the class centers of the plurality of influence classes, wherein the influence classes are obtained based on clustering of a plurality of sample factor vectors;

[0011] determining a load warning level according to the target class center, a class radius of the influence class, and the distance between the factor vector and the target class center.

[0012] In a possible implementation, the feature extraction and weighting processing on the plurality of factor data sets to obtain a factor vector includes:

[0013] obtaining a plurality of feature filtering vectors corresponding to the plurality of factor data sets and weighting values, wherein the weighting values are determined based on the influence of the features of the factor data sets on the tower body load;

[0014] performing feature extraction on the plurality of factor data sets by using the plurality of feature filtering vectors to obtain a plurality of feature vectors;

[0015] performing pooling processing on the plurality of feature vectors respectively to obtain a plurality of feature values corresponding to the plurality of feature vectors;

[0016] performing weighting processing and arrangement on the plurality of feature values according to the corresponding relationship between the plurality of feature values and the plurality of factor data sets to obtain a factor vector.

[0017] In a possible implementation, the feature extraction on the plurality of factor data sets by using the plurality of feature filtering vectors to obtain a plurality of feature vectors includes:

[0018] For each factor data set, the following steps are performed:

[0019] initializing an extraction position;

[0020] intermediate vector extraction step: extracting a plurality of data with the same dimension as the feature filtering vector from the factor data set according to the extraction position as an intermediate vector;

[0021] calculating the vector product of the feature filtering vector and the intermediate vector;

[0022] adding the vector product to the feature vector;

[0023] if the intermediate vector does not include the last data in the factor data set, moving the extraction position and jumping to the intermediate vector extraction step.

[0024] In a possible implementation, the weighting values are determined based on a tower body load linear model, including:

[0025] obtaining a plurality of tower body load sample values, a plurality of factor vector samples corresponding to the plurality of tower body load sample values, and a linear model, wherein the linear model is:

[0026]

[0027] wherein Load is a tower body load sample, wn is an nth weighted value, Svector(n) is an nth element of a factor vector sample, and N is a total number of elements of the factor vector sample;

[0028] solving the linear model according to the plurality of tower body load sample values and the plurality of factor vector samples to determine a plurality of weighted values.

[0029] In a possible implementation, a characteristic filtering vector corresponding to a factor dataset is determined based on an influence analysis of sample factor data on sample tower body load, and includes:

[0030] obtaining a sample tower body load sample dataset, a sample factor dataset corresponding to the sample tower body load sample dataset, and a plurality of undetermined characteristic filtering vectors, wherein the plurality of undetermined characteristic filtering vectors include at least one of the following: an accumulated characteristic extraction vector, a differential characteristic extraction vector, and a curvature characteristic extraction vector;

[0031] performing characteristic extraction on the sample factor dataset according to the plurality of undetermined characteristic filtering vectors to obtain a plurality of undetermined characteristic vectors corresponding to the plurality of undetermined characteristic filtering vectors;

[0032] determining a plurality of correlation coefficients corresponding to the plurality of undetermined characteristic vectors according to the plurality of undetermined characteristic vectors and the sample tower body load sample dataset;

[0033] selecting a coefficient with a largest absolute value from the plurality of correlation coefficients as a target coefficient;

[0034] selecting an undetermined characteristic filtering vector corresponding to the target coefficient as a characteristic filtering vector.

[0035] In a possible implementation, the determining of the load early warning level according to the target class center, a class radius of an influence class, and a distance between the factor vector and the target class center includes:

[0036] obtaining an early warning level of a class to which the target class center belongs as a target early warning level, wherein the early warning level is determined based on a class center;

[0037] if the target early warning level is a lowest level in a plurality of influence classes, and the distance between the factor vector and the target class center is greater than a class radius of the influence class, setting the load early warning level to the lowest level;

[0038] if the target warning level is the highest level in the plurality of influence classes, and the distance between the factor vector and the target class center is greater than the class radius of the influence class, setting the load warning level to the highest level;

[0039] Otherwise, setting the load warning level to the target warning level.

[0040] In a possible implementation, the warning level is determined based on the class center, comprising:

[0041] obtaining a tower load linear model and a plurality of class centers corresponding to the plurality of influence classes;

[0042] inputting the plurality of class centers into the tower load linear model respectively to obtain a plurality of output values of the model;

[0043] sorting the plurality of influence classes according to the sizes of the plurality of output values;

[0044] taking the sequence number obtained by the sorting as the warning levels of the plurality of influence classes.

[0045] In a second aspect, an embodiment of the present application provides a tower load warning device for implementing the tower load warning method in the first aspect or any possible implementation manner of the first aspect, and the tower load warning device comprises:

[0046] a factor data acquisition module, configured to acquire a plurality of factor data sets, wherein the plurality of factor data sets are acquired based on the same time period, each factor data set comprises a plurality of factor data corresponding to a plurality of time nodes, and the factor data affects the tower load;

[0047] a factor vector extraction module, configured to perform feature extraction and weighting processing on the plurality of factor data sets to obtain a factor vector;

[0048] a target class center determination module, configured to determine, from class centers of the plurality of influence classes, a class center with the smallest distance from the factor vector as a target class center, wherein the influence classes are obtained by clustering a plurality of sample factor vectors;

[0049] and,

[0050] a load warning level determination module, configured to determine a load warning level according to the target class center, a class radius of the influence class, and a distance between the factor vector and the target class center.

[0051] In a third aspect, an embodiment of the present application provides a terminal, comprising a memory and a processor, wherein the memory stores a computer program capable of running on the processor, and the processor implements the steps of the method according to the first aspect or any possible implementation manner of the first aspect when running the computer program.

[0052] In a fourth aspect, an embodiment of the present application provides a computer readable storage medium, which stores a computer program, and the computer program implements the steps of the method according to the first aspect or any possible implementation manner of the first aspect when executed by a processor.

[0053] Compared with the prior art, the embodiment of the present application has the following beneficial effects:

[0054] The embodiment of the present application discloses a tower load early warning method, which first acquires a plurality of factor data sets, wherein the plurality of factor data sets are acquired based on the same time period, and each factor data set includes a plurality of factor data corresponding to a plurality of time nodes, and the factor data affects the tower body load; then, the plurality of factor data sets are subjected to feature extraction and weighting processing to obtain a factor vector; then, a class center with the smallest distance from the factor vector is determined as a target class center from a plurality of influence class centers, wherein the influence class is obtained based on clustering of a plurality of sample factor vectors; finally, according to the target class center, a class radius of the influence class, and a distance between the factor vector and the target class center, a load early warning level is determined. The embodiment of the present application determines a feature filtering vector based on the factor data set by using a correlation analysis method, extracts useful features based on the filtering vector, constructs a linear load model based on the features, determines the coefficients of each feature through the load model, thereby constructing the factor vector, and finally determines the early warning level according to the belonging class through the classification method. Therefore, the early warning level division is more scientific and has more basis, thereby solving the problem that the risk amount caused by external factors is not easy to be determined. BRIEF DESCRIPTION OF DRAWINGS

[0055] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without any creative labor.

[0056] Figure 1 is a flowchart of the tower load early warning method provided by the embodiment of the present application;

[0057] Figure 2 is a principle diagram of the factor vector classification process provided by the embodiment of the present application;

[0058] Figure 3 is a functional block diagram of a tower load early warning device provided by an embodiment of the present application;

[0059] Figure 4 is a functional block diagram of a terminal provided by an embodiment of the present application. DETAILED DESCRIPTION

[0060] In the following description, for purposes of explanation and not limitation, specific details are set forth, such as a particular sequence of steps, in order to provide a thorough understanding of embodiments of the present application. However, it will be apparent to those skilled in the art that the present application can be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known methods, devices, and apparatus are omitted so as not to obscure the description of the present application with unnecessary detail.

[0061] In order to make the objects, technical solutions and advantages of the present application clearer, the following will be described in conjunction with the accompanying drawings and specific embodiments.

[0062] The following will be described in detail for the embodiments of the present application. The present example is implemented on the premise of the technical solutions of the present application, and detailed implementation manners and specific operation processes are given. However, the protection scope of the present application is not limited to the following embodiments.

[0063] Figure 1 is a flowchart of a tower load early warning method provided by an embodiment of the present application.

[0064] As shown in Figure 1 , it shows an implementation flowchart of a tower load early warning method provided by an embodiment of the present application, and the details are as follows:

[0065] In step 101, a plurality of factor data sets are acquired, wherein the plurality of factor data sets are acquired based on the same time period, each factor data set includes a plurality of factor data corresponding to a plurality of time nodes, and the factor data affects the tower body load.

[0066] In step 102, feature extraction and weighting processing are performed on the plurality of factor data sets to obtain a factor vector.

[0067] In some embodiments, the step 102 includes:

[0068] a plurality of feature filtering vectors corresponding to the plurality of factor data sets and weighting values are acquired, wherein the weighting values are determined based on the influence of the features of the factor data set on the tower body load;

[0069] feature extraction is performed on the plurality of factor data sets through the plurality of feature filtering vectors to obtain a plurality of feature vectors;

[0070] performing pooling processing on the plurality of feature vectors respectively to obtain a plurality of feature values corresponding to the plurality of feature vectors;

[0071] performing weighting processing and arrangement on the plurality of feature values according to a corresponding relationship between the plurality of feature values and the plurality of factor data sets to obtain a factor vector.

[0072] In some embodiments, the feature extraction on the plurality of factor data sets to obtain a plurality of feature vectors through the plurality of feature filtering vectors comprises:

[0073] For each factor data set, the following steps are performed:

[0074] initializing an extraction position;

[0075] intermediate vector extraction step: extracting a plurality of data with the same dimension as the feature filtering vector from the factor data set according to the extraction position as an intermediate vector;

[0076] calculating the vector product of the feature filtering vector and the intermediate vector;

[0077] adding the vector product to the feature vector;

[0078] If the intermediate vector does not include the last data in the factor data set, the extraction position is moved, and the intermediate vector extraction step is jumped to.

[0079] In some embodiments, the weighting values are determined based on a tower body load linear model, comprising:

[0080] obtaining a plurality of tower body load sample values, a plurality of factor vector samples corresponding to the plurality of tower body load sample values, and a linear model, wherein the linear model is:

[0081]

[0082] wherein Load is the tower body load sample, wn is the nth weighting value, Svector(n) is the nth element of the factor vector sample, and N is the total number of elements of the factor vector sample;

[0083] solving the linear model according to the plurality of tower body load sample values and the plurality of factor vector samples to determine a plurality of weighting values.

[0084] In some embodiments, the feature filtering vector corresponding to the factor data set is determined based on the influence analysis of the sample factor data on the sample tower body load, comprising:

[0085] Obtaining a sample tower body load sample data set, a sample factor data set corresponding to the sample tower body load sample data set, and a plurality of to-be-determined characteristic filtering vectors, wherein the plurality of to-be-determined characteristic filtering vectors include at least one of the following: an accumulated characteristic extraction vector, a differential characteristic extraction vector, and a curvature characteristic extraction vector;

[0086] Characteristic extraction is performed on the sample factor data set according to the plurality of to-be-determined characteristic filtering vectors, and a plurality of to-be-determined characteristic vectors corresponding to the plurality of to-be-determined characteristic filtering vectors are obtained.

[0087] According to the plurality of to-be-determined characteristic vectors and the sample tower body load sample data set, a plurality of correlation coefficients corresponding to the plurality of to-be-determined characteristic vectors are determined.

[0088] The coefficient with the largest absolute value is selected as a target coefficient from the plurality of correlation coefficients.

[0089] The to-be-determined characteristic filtering vector corresponding to the target coefficient is selected as a characteristic filtering vector.

[0090] Exemplarily, a factor refers to a characteristic that affects the health state of a tower, for example, a wind speed, an air temperature, snowfall, and the like. When air humidity is large or precipitation occurs, the air temperature often drops sharply, and ice load often occurs on a power line. When the ice load on both sides of the tower is uneven, there is a risk of tower load.

[0091] Therefore, it is necessary to analyze the load of the tower in combination with the factors. In fact, the influence of factors on the load of the tower is multifaceted. Some factors are the result of long-term effects, for example, the effect of wind on the tower. Some factors only affect the health state of the tower when they change sharply, for example, the air temperature. Based on this, the embodiment of the present application adopts a characteristic extraction method to extract characteristics. Specifically, a filtering vector is used to perform a characteristic extraction operation on a factor data set through convolution calculation. The filtering vector can have various forms, for example, [-1 / n, 1 / n, …, 1 / n] is used to extract the mean value characteristic of n data, [1, -2, 1] is used to extract the second-order differential characteristic, and [1, 1, …, 1] is used to extract the characteristic of the accumulation of multiple data.

[0092] After the characteristics are extracted, a maximum pooling processing or a mean pooling processing is performed, that is, the maximum value or the mean value is selected from a plurality of elements of a characteristic vector as a characteristic value. The characteristic values corresponding to a plurality of characteristic vectors are arranged to form a factor vector. These factor vectors are actually the extraction of key characteristics of factor data.

[0093] As mentioned above, the feature filtering vector actually adopted by different factor data is different, because the principle of different factor data affecting the load of the tower is different. The embodiment of the present application is based on the correlation analysis after the data is extracted by using different filtering vectors, and the feature filtering vector is determined according to the result of the correlation analysis (for example, the correlation analysis is performed by using the Spearman correlation coefficient).

[0094] As for the weight aspect, the present application constructs a linear equation for expressing the relationship between the above factor vector and the tower body load, and the linear equation expression is as follows:

[0095]

[0096] In the formula, Load is the tower body load sample, wn is the nth weighted value, Svector(n) is the nth element of the factor vector sample, and N is the total number of elements of the factor vector sample.

[0097] In fact, the weight is the weight in the linear equation, and as for the solution of the weight, the embodiment of the present application obtains a plurality of tower body load sample values, and a plurality of factor vector samples corresponding to the plurality of load sample values, and brings the plurality of tower body load sample values and the plurality of factor vector samples into the above linear equation respectively, so as to obtain an equation group about the plurality of weighted values. The equation group is solved, and the solution of the weighted value is obtained.

[0098] In step 103, a class center with the smallest distance from the factor vector is determined as a target class center from a plurality of class centers of influence classes, wherein the influence classes are obtained based on clustering of a plurality of sample factor vectors.

[0099] In step 104, a load warning level is determined according to the target class center, the class radius of the influence class, and the distance between the factor vector and the target class center.

[0100] In some embodiments, the step 104 includes:

[0101] A warning level of a class to which the target class center belongs is obtained as a target warning level, wherein the warning level is determined based on the class center;

[0102] If the target warning level is the lowest level in the plurality of influence classes, and the distance between the factor vector and the target class center is greater than the class radius of the influence class, the load warning level is set to the lowest level;

[0103] If the target warning level is the highest level in the plurality of influence classes, and the distance between the factor vector and the target class center is greater than the class radius of the influence class, the load warning level is set to the highest level;

[0104] Otherwise, set the load warning level as the target warning level.

[0105] In some embodiments, the warning level is determined based on the class center, comprising:

[0106] Obtaining a tower body load linear model and a plurality of class centers corresponding to the plurality of influence classes;

[0107] Inputting the plurality of class centers into the tower body load linear model respectively to obtain a plurality of output values of the model;

[0108] According to the size of the plurality of output values, sorting the plurality of influence classes;

[0109] Taking the sequence number obtained by sorting as the warning level of the plurality of influence classes.

[0110] Exemplarily, in the embodiments of the present application, the load warning level is determined by judging the distance of the factor vector to the class center of the classified classes, for example, using the Euclidean distance judgment method, finding a nearest class center, and determining the load warning level according to the warning level of the class center.

[0111] In fact, before the distance judgment and the finding of the target class center, we have already performed clustering and completed the classification operation through the existing samples. There are various clustering methods, and in fact, in order to facilitate the judgment of the warning level later and improve the accuracy of the judgment, the embodiments of the present application recommend using the k-means clustering method. This method usually specifies a plurality of random clustering centers, and the remaining samples are classified into the clustering centers according to the nearest distance principle. After the clustering is completed, a new clustering center is selected for each class through distance calculation. If the distance between the new clustering center and the original clustering center is small, the clustering is completed. Otherwise, the new clustering center is used again to cluster according to the nearest distance principle, and the iteration is repeated until the clustering is completed. When the distribution quantity of the plurality of samples in the class clustered by using this clustering method is large, it generally conforms to the spherical distribution.

[0112] Figure 2 The process of finding the target class center according to the factor vector clustering center through distance judgment is shown. In the figure, the distance of the factor vector 201 to the class center 203 of each class 202 is calculated, and the nearest class center 203 is selected as the target class center. Obviously, if the distance of the factor vector 201 to the class center is less than the class radius R, the factor vector 201 can be classified into the class, and the load warning level is the warning level of the class where the target class center is located. If the distance of the factor vector 201 to the class center is greater than the class radius R, it means that the vector exceeds the existing class, and only the distance to the target class center is relatively close. In this case, the target class center will become a reference vector.

[0113] At this time, if the warning level of the class to which the target class center belongs is the lowest level of the influence class, it indicates that the factor vector 201 is lower than the existing warning level, and the load warning level should be the lowest level; and if the warning level of the class to which the target class center belongs is the highest level of the influence class, it indicates that the factor vector 201 is higher than the existing warning level, and the load warning level should be the highest level.

[0114] Here, there is actually a problem of how each class determines its warning level. One way provided by the present application is to input the class center into the tower load linear model as described above, obtain the result of the model, and the meaning represented by the output result of the model is the estimated value of the tower load according to the class center. The higher the value, the greater the risk, and the warning level should also be higher accordingly. When multiple class centers are input into the model to obtain multiple estimated values, the estimated values are sorted, and the warning level is determined according to the sorted values.

[0115] The tower load warning method embodiment of the present application first obtains a plurality of factor data sets, wherein the plurality of factor data sets are obtained based on the same time period, and each factor data set includes a plurality of factor data corresponding to a plurality of time nodes, and the factor data affects the tower load. Then, feature extraction and weighting processing are performed on the plurality of factor data sets to obtain a factor vector. Next, a class center with the smallest distance from the factor vector is determined as a target class center from among a plurality of influence class centers, wherein the influence class is obtained based on clustering of a plurality of sample factor vectors. Finally, a load warning level is determined according to the target class center, a class radius of the influence class, and a distance between the factor vector and the target class center. The embodiment of the present application determines a feature filtering vector based on the correlation analysis method of the factor data set, extracts useful features based on the filtering vector, and constructs a linear load model based on the features. The coefficients of each feature are determined through the load model to construct a factor vector. Finally, the warning level is determined according to the belonging class through the classification method. Therefore, the warning level division is more scientific and has more basis, thereby solving the problem that the risk amount caused by external factors is not easy to determine.

[0116] It should be understood that the size of the serial number of each step in the above embodiment does not mean the order of execution, and the execution order of each process should be determined according to its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present application.

[0117] The following is a device embodiment of the present application. For details not described in detail, please refer to the corresponding method embodiments described above.

[0118] Figure 3 is a functional block diagram of the tower load warning device provided by the embodiment of the present application. Referring to Figure 3The tower load early warning device 3 comprises a factor data acquisition module 301, a factor vector extraction module 302, a target class center determination module 303, and a load early warning level determination module 304, wherein:

[0119] The factor data acquisition module 301 is configured to acquire a plurality of factor data sets, wherein the plurality of factor data sets are acquired based on the same time period, each factor data set comprises a plurality of factor data corresponding to a plurality of time nodes, and the factor data influences the tower load;

[0120] The factor vector extraction module 302 is configured to perform feature extraction and weighting processing on the plurality of factor data sets to obtain a factor vector;

[0121] The target class center determination module 303 is configured to determine, from class centers of a plurality of influence classes, a class center with the smallest distance from the factor vector as a target class center, wherein the influence classes are obtained based on clustering of a plurality of sample factor vectors;

[0122] The load early warning level determination module 304 is configured to determine a load early warning level according to the target class center, a class radius of the influence class, and a distance between the factor vector and the target class center.

[0123] Figure 4 is a functional block diagram of a terminal provided by an embodiment of the present application. As shown in Figure 4 The terminal 4 of the embodiment comprises a processor 400 and a memory 401, and the memory 401 stores a computer program 402 which can run on the processor 400. The processor 400 implements the steps in the above-described various tower load early warning methods and embodiments when executing the computer program 402, such as Figure 1 Steps 101 to 104 shown in FIG. 1.

[0124] For example, the computer program 402 can be divided into one or more modules / units, which are stored in the memory 401 and executed by the processor 400 to complete the present application.

[0125] The terminal 4 can be a desktop computer, a notebook computer, a palm computer, a cloud server, and the like. The terminal 4 can include, but is not limited to, the processor 400 and the memory 401. Those skilled in the art can understand that Figure 4 The terminal 4 is only an example and does not constitute a limitation on the terminal 4, and can include more or fewer components than shown, or combine certain components, or different components, for example, the terminal 4 can also include an input / output device, a network access device, a bus, and the like.

[0126] The processor 400 can be a central processing unit (CPU), and can also be other general-purpose processors, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or the like. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0127] The memory 401 can be an internal storage unit of the terminal 4, such as a hard disk or a memory of the terminal 4. The memory 401 can also be an external storage device of the terminal 4, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, or the like, which is equipped on the terminal 4. Further, the memory 401 can also include both the internal storage unit and the external storage device of the terminal 4. The memory 401 is used to store the computer program 402 and other programs and data required by the terminal 4. The memory 401 can also be used to temporarily store data that has been output or is to be output.

[0128] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above functional units and modules is exemplified, and in actual application, the above functions can be completed by different functional units and modules according to needs, that is, the internal structure of the apparatus is divided into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiments can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit, and the integrated unit can be realized in the form of hardware or in the form of a software functional unit. In addition, the specific names of each functional unit and module are only for the purpose of mutual distinction, and do not limit the protection scope of the present application. The specific working process of the units and modules in the system can refer to the corresponding process in the foregoing method embodiments, which will not be described here.

[0129] In the above embodiments, the description of each embodiment has its own emphasis, and the parts not described or recorded in detail in a certain embodiment can refer to the relevant description of other embodiments.

[0130] Those skilled in the art can understand that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0131] In the embodiments provided by the present application, it should be understood that the disclosed apparatus / terminal and method can be implemented in other manners. For example, the described apparatus / terminal embodiments are merely schematic, and the division of the modules or units can be different, for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections can be indirect couplings or communication connections through some interfaces, devices or units, and can be in electrical, mechanical or other forms.

[0132] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, i.e., can be located in one place, or can be distributed on multiple network units. Some or all of the units can be selected according to actual needs to achieve the purposes of the embodiments.

[0133] In addition, each functional unit in the various embodiments of the present application can be integrated into one processing unit, or each unit can exist physically as a separate unit, or two or more units can be integrated into one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0134] The integrated module / unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, all or part of the processes in the above-mentioned embodiment methods can also be completed by a computer program instructing related hardware, and the computer program can be stored in a computer readable storage medium. When the computer program is executed by a processor, the steps of the above-mentioned various method and device embodiments can be implemented. The computer program includes computer program code, which can be in the form of source code, object code, executable files or some intermediate forms. The computer readable medium can include any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc.

[0135] The above-mentioned embodiments are only used to illustrate the technical solutions of the present application, rather than limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacements for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A method for early warning of tower loads, characterized in that, include: Multiple factor datasets are acquired, wherein the factor datasets are acquired based on the same time period, and the factor datasets include multiple factor data corresponding to multiple time nodes, and the factor data affect the tower load; Feature extraction and weighting are performed on the multiple factor datasets to obtain factor vectors, including: Obtain multiple feature filtering vectors and weighted values ​​corresponding to the multiple factor datasets, wherein the weighted values ​​are determined based on the influence of the features of the factor datasets on the tower load; By using the multiple feature filtering vectors, feature extraction is performed on the multiple factor datasets to obtain multiple feature vectors; Pooling is performed on the multiple feature vectors respectively to obtain multiple feature values ​​corresponding to the multiple feature vectors; Based on the correspondence between the multiple feature values ​​and the multiple factor datasets, the multiple feature values ​​are weighted and arranged to obtain a factor vector; The target class center is determined from the class centers of the multiple influence classes that have the smallest distance to the factor vector. The influence classes are obtained based on the clustering of multiple sample factor vectors. The load warning level is determined based on the target class center, the class radius of the influencing class, and the distance between the factor vector and the target class center; The step of extracting features from the multiple factor datasets using the multiple feature filtering vectors to obtain multiple feature vectors includes: For each factor dataset, perform the following steps: Initialize the extraction location; Intermediate vector extraction step: Based on the extraction position, extract multiple data points from the factor dataset with the same dimension as the feature filter vector as intermediate vectors; Calculate the vector product of the feature filtering vector and the intermediate vector; Add the vector product to the feature vector; If the intermediate vector does not include the last data in the factor dataset, then the extraction position is moved and the process jumps to the intermediate vector extraction step. The feature filtering vector corresponding to the factor dataset is determined based on the influence analysis of the sample factor data on the sample tower load, including: Obtain a sample tower load sample dataset, a sample factor dataset corresponding to the sample tower load sample dataset, and multiple undetermined feature filtering vectors, wherein the multiple undetermined feature filtering vectors include at least one of the following: cumulative feature extraction vector, differential feature extraction vector, and curvature feature extraction vector; Based on the multiple undetermined feature filtering vectors, feature extraction is performed on the sample factor dataset to obtain multiple undetermined feature vectors corresponding to the multiple undetermined feature filtering vectors; Based on the plurality of undetermined feature vectors and the sample tower load sample dataset, determine a plurality of correlation coefficients corresponding to the plurality of undetermined feature vectors; Select the coefficient with the largest absolute value from among the multiple correlation coefficients as the target coefficient; Select the undetermined feature overfilter vector corresponding to the target coefficient as the feature filtering vector.

2. The tower load early warning method according to claim 1, characterized in that, The weighted values ​​are determined based on a linear model of tower load, including: Obtain multiple tower load sample values, multiple factor vector samples corresponding to the multiple tower load sample values, and a linear model, wherein the linear model is: In the formula, For tower load samples, For the first Each weighted value, For the factor vector sample One element, This represents the total number of elements in the factor vector sample. The linear model is solved based on the multiple tower load sample values ​​and the multiple factor vector samples to determine multiple weighting values.

3. The tower load early warning method according to any one of claims 1-2, characterized in that, The step of determining the load warning level based on the target class center, the class radius of the influencing class, and the distance between the factor vector and the target class center includes: The warning level of the class to which the target class center belongs is obtained as the target warning level, wherein the warning level is determined based on the class center; If the target warning level is the lowest level among multiple impact classes, and the distance between the factor vector and the center of the target class is greater than the class radius of the impact class, then the load warning level is set to the lowest level. If the target warning level is the highest level among multiple impact classes, and the distance between the factor vector and the center of the target class is greater than the class radius of the impact class, then the load warning level is set to the highest level. Otherwise, the load warning level will be set to the target warning level.

4. The tower load early warning method according to claim 3, characterized in that, The warning level is determined based on class centering, including: Obtain the linear model of the tower load and the corresponding class centers of the multiple influence classes; The multiple class centers are respectively input into the linear model of tower load to obtain multiple output values ​​of the model; The multiple influence classes are sorted according to the magnitude of the multiple output values; The sorted sequence number will be used as the warning level for the multiple impact categories.

5. A tower load early warning device, characterized in that, For implementing the tower load early warning method as described in any one of claims 1-4, the tower load early warning device comprises: The factor data acquisition module is used to acquire multiple factor datasets, wherein the factor datasets are acquired based on the same time period, and the factor datasets include multiple factor data corresponding to multiple time nodes, and the factor data affect the tower load. The factor vector extraction module is used to perform feature extraction and weighting processing on the multiple factor datasets to obtain factor vectors, including: Obtain multiple feature filtering vectors and weighted values ​​corresponding to the multiple factor datasets, wherein the weighted values ​​are determined based on the influence of the features of the factor datasets on the tower load; extract features from the multiple factor datasets using the multiple feature filtering vectors to obtain multiple feature vectors; perform pooling processing on the multiple feature vectors to obtain multiple feature values ​​corresponding to the multiple feature vectors; according to the correspondence between the multiple feature values ​​and the multiple factor datasets, perform weighted processing on the multiple feature values ​​and arrange them to obtain factor vectors; The target class center determination module is used to determine the class center with the smallest distance to the factor vector from the class centers of the multiple influence classes as the target class center, wherein the influence classes are obtained based on the clustering of multiple sample factor vectors; as well as, The load warning level determination module is used to determine the load warning level based on the target class center, the class radius of the influencing class, and the distance between the factor vector and the target class center; The step of extracting features from the multiple factor datasets using the multiple feature filtering vectors to obtain multiple feature vectors includes: For each factor dataset, perform the following steps: Initialize the extraction location; Intermediate vector extraction step: Based on the extraction position, extract multiple data points from the factor dataset with the same dimension as the feature filter vector as intermediate vectors; Calculate the vector product of the feature filtering vector and the intermediate vector; Add the vector product to the feature vector; If the intermediate vector does not include the last data in the factor dataset, then the extraction position is moved and the process jumps to the intermediate vector extraction step. The feature filtering vector corresponding to the factor dataset is determined based on the influence analysis of the sample factor data on the sample tower load, including: Obtain a sample tower load sample dataset, a sample factor dataset corresponding to the sample tower load sample dataset, and multiple undetermined feature filtering vectors, wherein the multiple undetermined feature filtering vectors include at least one of the following: cumulative feature extraction vector, differential feature extraction vector, and curvature feature extraction vector; Based on the multiple undetermined feature filtering vectors, feature extraction is performed on the sample factor dataset to obtain multiple undetermined feature vectors corresponding to the multiple undetermined feature filtering vectors; Based on the plurality of undetermined feature vectors and the sample tower load sample dataset, determine a plurality of correlation coefficients corresponding to the plurality of undetermined feature vectors; Select the coefficient with the largest absolute value from among the multiple correlation coefficients as the target coefficient; Select the undetermined feature overfilter vector corresponding to the target coefficient as the feature filtering vector.

6. A terminal, comprising a memory and a processor, wherein the memory stores a computer program executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 4 above.

7. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 4 above.

Images