Power transmission line engineering index mutual influence prediction method based on multi-dimensional data analysis

Abstract

The present application relates to the technical field of power index prediction, in particular to a power transmission line engineering index mutual influence prediction method based on multi-dimensional data analysis. The present application firstly obtains the time sequence of state parameters of each tower under each index and the current state vector of each tower; then obtains the index importance of each index, the state fluctuation coefficient of each tower and the state influence weight of each index of each tower, and further obtains the final state influence weight of each index of each tower at the current time, so as to obtain the predicted state vector of each tower in combination with the current state vector. The present application analyzes the change correlation and historical fluctuation of the state parameters under the multi-dimensional index of the tower, evaluates the attention weight of each index of each tower, adjusts the input of the Kalman filtering algorithm, makes the algorithm capture the index characteristics that have a great influence on the state change of the tower, and further improves the prediction accuracy of the state of each tower in the power transmission line engineering.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power index prediction, and particularly relates to a power transmission line engineering index mutual influence prediction method based on multi-dimensional data analysis. BACKGROUND

[0002] In a modern power system, a power transmission line is an important channel for transmitting power from a power plant to a user, and a tower is an important supporting structure for erecting the power transmission line, which will directly affect the reliability and stability of the power transmission line engineering; however, factors such as a changeable meteorological environment and a complex terrain condition will threaten the performance and safety of the power transmission line engineering, so safety monitoring and risk prediction of the power transmission line engineering are very important.

[0003] In the safety monitoring and risk prediction process of the power transmission line engineering, the current state of the tower and the power transmission line is evaluated by monitoring and analyzing state parameters under multi-dimensional indexes of the power transmission line engineering, and the current state is input into a Kalman filter to predict a future state; however, state parameters between the multi-dimensional indexes in the power transmission line engineering may have certain correlated changes, and each dimensional index has different influences on the state of the power transmission line engineering, so directly using the Kalman filter will affect the prediction accuracy of the future state of the power transmission line engineering. SUMMARY

[0004] In order to solve the technical problem of poor prediction accuracy of the future state of the power transmission line engineering in the prior art, the purpose of the present application is to provide a power transmission line engineering index mutual influence prediction method based on multi-dimensional data analysis, and the technical solution adopted is as follows:

[0005] At a current time, time series sequences of state parameters of each tower in the power transmission line engineering under each index in a preset historical period are obtained, and a current state vector of each tower is obtained; the current state vector includes state parameters under each index;

[0006] According to the correlated changes between different indexes of all towers, an index importance of each index is obtained; according to fluctuation changes of the time series sequence of each tower under each index, a state fluctuation coefficient of each tower is obtained relative to deviation of fluctuation changes of the time series sequence of other towers in a local geographical area under each index;

[0007] According to the fluctuation change of the state parameter in the time sequence under each index of each tower, combined with the state fluctuation coefficient of each tower and the index importance of each index, the state influence weight of each index of each tower is obtained; at the current time, according to the state influence weight of each index of each tower and the deviation of the state parameter of each tower under the same index from the state parameter of all towers, the final state influence weight of each index of each tower is obtained;

[0008] According to the final state influence weight and the current state vector, a predicted state vector of each tower at the next collection time is obtained.

[0009] Further, the index importance obtaining method comprises:

[0010] The state vector of each tower at each collection time is obtained, the state transition matrix of each tower is obtained based on the Kalman filter and the state vectors at all adjacent collection times, and the index importance of each index is obtained by synthesizing the matrix elements in the state transition matrix of all towers.

[0011] Further, the method for obtaining the index importance of each index by synthesizing the matrix elements in the state transition matrix of all towers comprises:

[0012] Each row and each column in the state transition matrix corresponds to an index; the sum of all elements in each row in the state transition matrix is taken as a first dependent reference value under the corresponding index, and the sum of all elements in each column in the state transition matrix is taken as a second dependent reference value under the corresponding index.

[0013] Under each index, the sum of the products of the first dependent reference value and the second dependent reference value of all towers is taken as a first parameter, the product of the sum of the first dependent reference value of all towers and the sum of the first dependent reference value of all towers is taken as a second parameter, the square sum of the second dependent reference value of all towers is taken as a third parameter, and the square sum of the sum of the second dependent reference value of all towers is taken as a fourth parameter.

[0014] The difference between the first parameter and the second parameter is taken as the numerator, the difference between the third parameter and the fourth parameter is taken as the denominator, and the fractional ratio is taken as the index importance of each index.

[0015] Further, the state fluctuation coefficient obtaining method comprises:

[0016] According to the fluctuation change of the state parameters in the time sequence of each tower under each index, a fluctuation evaluation parameter of each index of each tower is obtained; and the fluctuation evaluation parameters of all indexes of each tower are comprehensively obtained to obtain an overall fluctuation parameter of each tower.

[0017] Based on the connection of the towers in the power transmission line project, a tower sorting sequence in the power transmission line project is obtained; a fluctuation sequence is constructed according to the overall fluctuation parameter of each tower, and the sequence order of the fluctuation sequence is the same as that of the tower sorting sequence; and the tower sorting sequence is divided according to the change of the fluctuation sequence to obtain all tower subsequences.

[0018] According to the deviation degree of the overall fluctuation parameter of each tower from the overall fluctuation parameters of all towers in the tower subsequence to which the tower belongs, a state fluctuation coefficient of each tower is obtained; and the deviation degree is positively correlated with the state fluctuation coefficient.

[0019] Further, the method for obtaining the fluctuation evaluation parameter comprises:

[0020] According to the local fluctuation similarity of the time sequence of each tower under each index, the time sequence is segmented to obtain all sequence subsections.

[0021] According to the total number of the sequence subsections and the sequence length of the last sequence subsection in time sequence, a fluctuation evaluation parameter of each index of each tower is obtained; the sequence length is negatively correlated with the fluctuation evaluation parameter, and the total number is positively correlated with the fluctuation evaluation parameter.

[0022] Further, the method for obtaining the tower subsequence comprises:

[0023] A first-order difference sequence of the fluctuation sequence is obtained, if the adjacent two sequence elements in the first-order difference sequence are of opposite signs, then the maximum sequence number in the sequence numbers of the corresponding adjacent two sequence elements is taken as a cutting sequence number;

[0024] The sequence element corresponding to the cutting sequence number in the tower sorting sequence is taken as a sequence right end point of each tower subsequence, and the adjacent next sequence element of each sequence right end point is taken as a sequence left end point of the next tower subsequence.

[0025] A tower sequence subsection is obtained based on the sequence right end point and the sequence left end point of each tower subsequence.

[0026] Further, the method for obtaining the state influence weight comprises:

[0027] The state fluctuation coefficient of each tower is multiplied by the index importance of each index, and the product is taken as a first weight parameter; the fluctuation evaluation parameter of each index of each tower is normalized, and the normalized result is taken as a second weight parameter; the product of the first weight parameter and the second weight parameter is taken as the state influence weight of the corresponding index of the corresponding tower.

[0028] Further, the method for obtaining the final state influence weight comprises:

[0029] Under the same index, the fluctuation weight of each index of each tower is obtained according to the deviation of the state parameter of each tower from the average level of the state parameters of all towers; the state influence weight corresponding to the fluctuation weight is weighted, and the weighted result is taken as the final state influence weight of each index of each tower.

[0030] Further, the method for obtaining the predicted state vector comprises:

[0031] The current state vector of the corresponding tower is weighted by using the final state influence weight, the weighted result is taken as the input of the Kalman filter, and the predicted state vector of the corresponding tower at the next acquisition time is output.

[0032] Further, the index at least comprises load, tension ratio, height, span, weight, and amplitude of dancing.

[0033] The present application has the following beneficial effects:

[0034] The application obtains the time sequence of the state parameters of each tower in each index in the power transmission line project in a preset historical period and the current state vector of each tower at the current time, to provide a data basis for subsequent analysis of fluctuation correlation and prediction; according to the change correlation between different indexes of different towers, the index importance of each index is obtained; the index importance reflects the influence of the index on the state change of the power transmission line project to some extent, so as to facilitate subsequent determination of the attention degree of each index in the current state vector; according to the fluctuation change of the time sequence of each tower in each index, the deviation of the fluctuation change of the time sequence of each tower in each index from other towers in the local geographical area, the state fluctuation coefficient of each tower is obtained, which reflects the state stability of each tower and prepares for subsequent evaluation of the consideration weight of each index of each tower; according to the fluctuation change of the state parameters in the time sequence of each tower in each index, combined with the state fluctuation coefficient of each tower and the index importance of each index, the state influence weight of each index of each tower is obtained; at the current time, according to the state influence weight of each index of each tower and the deviation of the state parameters of each tower in the same index from the state parameters of all towers, the final state influence weight of each index of each tower is obtained; the state influence weight reflects the attention weight of each index of each tower, so as to capture the state change of the tower in time; the final state influence weight combines the state deviation degree of the tower at the current time, more accurately evaluates the attention degree of the state parameters of each index of each tower, and improves the state prediction accuracy; according to the final state influence weight and the current state vector, the predicted state vector of each tower at the next collection time is obtained. The application analyzes the change correlation and historical fluctuation of the state parameters under the multi-dimensional index of the tower, evaluates the attention weight of each index of each tower, adjusts the input of the Kalman filtering algorithm, makes the algorithm capture the index characteristics with great influence on the state change of the tower, and further improves the prediction accuracy of the state of each tower in the power transmission line project. BRIEF DESCRIPTION OF DRAWINGS

[0035] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, and the advantages thereof, a brief introduction will be given to the drawings needed in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can be obtained without creative labor based on these drawings.

[0036] Figure 1 A flowchart of a power transmission line project index mutual influence prediction method based on multi-dimensional data analysis provided by an embodiment of the present application;

[0037] Figure 2A schematic diagram of a dependence degree table provided by one embodiment of the present application;

[0038] Figure 3 A flow chart of a state fluctuation coefficient acquisition method provided by one embodiment of the present application. DETAILED DESCRIPTION

[0039] In order to further clarify the technical means and effects taken by the present application to achieve the predetermined object of the application, the following describes in detail the specific implementation, structure, features and effects of a power transmission line engineering index mutual influence prediction method based on multi-dimensional data analysis according to the present application, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0040] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0041] The following specifically describes the specific scheme of the power transmission line engineering index mutual influence prediction method based on multi-dimensional data analysis provided by the present application in combination with the accompanying drawings.

[0042] Please refer to Figure 1 which shows a flow chart of the power transmission line engineering index mutual influence prediction method based on multi-dimensional data analysis provided by one embodiment of the present application, specifically including:

[0043] Step S1, at the current time, acquire the time sequence sequence of the state parameters of each tower in each index in the power transmission line engineering in the preset historical period, and the current state vector of each tower; the current state vector includes the state parameters under each index.

[0044] In one embodiment of the present application, various sensors are set to collect the state parameters of each tower under the corresponding index at all collection time points in the preset historical period, and then all the state parameters of each tower under each index are sorted according to the collection time sequence to construct the time sequence sequence; wherein the preset historical period is set to one week of history at the current time, and the collection frequency is once per minute, and the implementer can also set it according to the actual needs; at the same time, the current state vector of each tower is acquired at the current time, wherein the current state vector is a multi-dimensional vector containing the state parameters of each tower under the multi-dimensional index at the current time.

[0045] It should be noted that all the state parameters need to be standardized before constructing the time sequence sequence to eliminate the dimension for subsequent analysis and application.

[0046] In one embodiment of the present application, the indexes at least include load, tension ratio, height, span, weight, and dancing amplitude; wherein the load is calculated by measuring the strain information of the tower in real time through the strain gauge installed on the tower; the tension ratio is obtained by measuring the cable tension in real time through the tension sensor installed on the transmission cable between the tower and the next adjacent tower; the height of each tower and the span between two adjacent towers are obtained by using the laser range finder; the weight of the tower is collected by using the gravity sensor installed on the tower base and the like; and the dancing amplitude of the tower is monitored in real time by using the acceleration sensor. It should be noted that the collection of the state parameters of the above indexes is the prior art well known to those skilled in the art, and will not be described here. In other embodiments, the implementer can also set other types or quantities of indexes, which can be the structure stability related indexes of the tower such as vibration and inclination, and can also be the foundation related indexes around the tower such as soil bearing capacity or compactness, and the like, which will not be described here.

[0047] In step S2, the index importance of each index is obtained according to the change correlation between different indexes of all towers; and the state fluctuation coefficient of each tower is obtained according to the fluctuation change of the time sequence of each tower under each index, and the deviation of the fluctuation change of the time sequence of other towers in the local geographical area under each index.

[0048] Considering that there is a certain change correlation between different indexes, for example, the greater the load is, the greater the dancing amplitude can be, the index importance of each index can be evaluated by comprehensively analyzing the change correlation between different indexes of all towers; the index importance reflects the influence of the change of each index on the change of other indexes, and reflects the influence on the state change of the transmission line project.

[0049] Preferably, in one embodiment of the present application, considering that the state transition matrix in the Kalman filter can be used to describe how the system state changes over time, and each element in the state transition matrix reflects the mutual relationship between different state variables and how they affect the evolution of the system, the correlation change between different indexes can be analyzed and evaluated based on the state transition matrix, so as to obtain the index importance of each index; based on this, the method for obtaining the index importance includes:

[0050] The state vector of each tower at each collection time is obtained, and the state transition matrix of each tower is obtained based on the Kalman filter and the state vectors at all adjacent collection times; the matrix elements in the state transition matrices of all towers are comprehensively analyzed to obtain the index importance of each index.

[0051] It should be noted that the method for obtaining the state vector of each tower at each acquisition time is consistent with the method for obtaining the current state vector; the state transition matrix of the Kalman filter for obtaining the state vector at the adjacent acquisition time is known to those skilled in the art as prior art, and will not be described here.

[0052] In a preferred embodiment of the present application, considering that the state vector is an n-dimensional vector, the size of the state transition matrix is n x n, wherein each matrix element such as Aij represents the degree of influence of the jth state vector on the ith state vector in time stepping; and considering that the state vector is composed of state parameters under multiple dimensions, the state transition matrix can be regarded as a degree of index dependence table between each tower corresponding to different indexes, each row and each column in the degree of index dependence table can be regarded as an index corresponding dimension, and each value therein reflects the change correlation degree of the state parameter under each index and the state parameter under another index; therefore, the method for obtaining the index importance of each index by comprehensively considering the matrix elements in the state transition matrix of all towers includes:

[0053] Each row and each column in the state transition matrix corresponds to an index; the sum of all elements in each row in the state transition matrix is taken as the first dependence reference value under the corresponding index, and the sum of all elements in each column in the state transition matrix is taken as the second dependence reference value under the corresponding index.

[0054] Under each index, the sum of the products of the first dependence reference value and the second dependence reference value of all towers is taken as a first parameter, the product of the sum of the first dependence reference values of all towers and the sum of the first dependence reference values of all towers is taken as a second parameter, the square sum of the second dependence reference values of all towers is taken as a third parameter, and the square sum of the sum of the second dependence reference values of all towers is taken as a fourth parameter.

[0055] The difference between the first parameter and the second parameter is taken as the numerator, the difference between the third parameter and the fourth parameter is taken as the denominator, and the fractional ratio is taken as the index importance of each index.

[0056] As an example, please participate Figure 2 which shows a schematic diagram of a degree of index dependence table provided by an embodiment of the present application, each row and each column in the degree of index dependence table corresponds to an index dimension; the sum of all elements in each row in the state transition matrix, i.e. the degree of index dependence table, is taken as the first dependence reference value under the corresponding index, and similarly, the sum of all elements in each column in the state transition matrix, i.e. the degree of index dependence table, is taken as the second dependence reference value under the corresponding index; the first dependence reference value and the second dependence reference value respectively evaluate the influence of the change of each index on the change of another index in the row and column angles; then the index importance of each index is obtained.

[0057] The calculation formula of the index importance is: ; wherein, i is the serial number of the index; a is the serial number of the tower; is the total number of the towers in the transmission line; is the index importance of the i-th index; is the first dependent reference value of the i-th index of the a-th tower; is the second dependent reference value of the i-th index of the a-th tower; is the first parameter; is the second parameter; is the third parameter; is the fourth parameter; is the absolute value symbol.

[0058] In the calculation formula of the index importance, the absolute value of the difference between the square sum of the second dependent reference values of all the towers under each index and the square sum of the first dependent reference values of all the towers under each index reflects the dispersion degree of the second dependent reference values of different towers under the index. The greater the dispersion degree, the more scattered the influence of the index on the changes of the remaining indexes, and the lower the importance of the index. The combination operation between the first dependent reference values and the second dependent reference values of all the towers under each index is used to evaluate the importance of each index. The greater the difference, the more likely the changes of the index significantly affect the data performance of the remaining indexes.

[0059] It should be noted that the possibility of the second dependent reference values of different towers being equal and 0 under each index is extremely low, so the case of the denominator being 0 is ignored. When the extreme case of the denominator being 0 exists, the implementer can set the denominator to a very small non-zero positive parameter such as 0.01, which makes the fraction meaningful without affecting the value of the fraction.

[0060] In another embodiment of the present application, the implementer can also obtain the correlation coefficient such as the Pearson correlation coefficient between each two time series under all the indexes of each tower to evaluate the change correlation between different indexes in each tower. Then, the average of the Pearson correlation coefficients between the corresponding time series of all the towers between each two indexes is taken as the correlation change coefficient between the two indexes. Taking any index as a target index, the correlation change coefficients between the target index and the remaining indexes are accumulated and then linearly normalized to obtain the index importance of the target index, and further to obtain the index importance of each index.

[0061] It should be noted that the Pearson correlation coefficient and linear normalization are well-known prior art to those skilled in the art, and will not be described here. In other embodiments, the implementer can also use other correlation coefficients and normalization methods.

[0062] Further, the fluctuation of the time sequence of each tower under each index reflects the state of each tower, and the more intense the fluctuation of the state parameter of the tower under the index, the more likely the tower is to change its stable state and cause a safety risk. Therefore, when predicting the future state of the tower based on the current state of the tower, more attention should be paid to the change of the state parameter of the index. Therefore, the embodiment of the present application further obtains the state fluctuation coefficient of each tower according to the fluctuation of the time sequence of each tower under each index and the deviation of the fluctuation of the time sequence of other towers in the local geographical area under each index.

[0063] Preferably, in an embodiment of the present application, the method for obtaining the state fluctuation coefficient comprises:

[0064] Referring to Figure 3 , a flowchart of a method for obtaining a state fluctuation coefficient according to an embodiment of the present application is shown, which specifically comprises:

[0065] In step S201, the fluctuation evaluation parameter of each index of each tower is obtained according to the fluctuation of the state parameter in the time sequence of each tower under each index. The overall fluctuation parameter of each tower is obtained by comprehensively considering the fluctuation evaluation parameters of all indexes of each tower.

[0066] The greater the fluctuation of the state parameter in the time sequence of each tower under each index, the greater the influence on the change of the stable state of the tower. If the state parameters of the tower under all indexes fluctuate greatly, the possibility of the change of the state of the tower is greater. Therefore, the overall fluctuation parameter of each tower is obtained by comprehensively considering the fluctuation of the time sequence of each tower under all indexes.

[0067] The overall fluctuation parameter reflects the fluctuation degree of each index of the tower in a preset historical period at the current time, and reflects the possibility of the change of each tower, which prepares for the subsequent analysis of the state influence weight of each index of each tower.

[0068] In a preferred embodiment of the present application, the fluctuation of the time sequence is not instantaneous mutation, but has a certain transition change over time. Therefore, the time sequence can be segmented according to the local fluctuation of the time sequence. The more the number of segments, the faster the fluctuation. Further, the longer the length of the local segment adjacent to the current time, the more stable the state of the time sequence adjacent to the current time. Therefore, the method for obtaining the fluctuation evaluation parameter comprises:

[0069] According to the local fluctuation similarity of the time sequence of each tower under each index, the time sequence is segmented to obtain all sequence segments;

[0070] According to the total number of sequence segments and the sequence length of the last sequence segment in time sequence, the fluctuation evaluation parameter of each tower under each index is obtained; the sequence length is negatively correlated with the fluctuation evaluation parameter, and the total number is positively correlated with the fluctuation evaluation parameter.

[0071] As an example, the time sequence of each tower under each index is specifically differentiated to obtain a differential time sequence, if the differential time sequence is curve fitted to obtain all extreme points in the curve, the extreme points are taken as the mutation points of fluctuation change, and the horizontal axis number of the extreme points in the curve is taken as the sequence segmentation point in the time sequence after being added by 1, wherein the sequence segmentation point is the starting point of each sequence segment, and the previous adjacent sequence element of the sequence segmentation point is the end point of the previous sequence segment; all sequence segments are determined based on the starting point and the end point of the sequence segment.

[0072] In another example, the time sequence of each tower under each index is specifically clustered, such as using a hierarchical clustering algorithm to obtain all clusters, and the state parameters in the same cluster are assigned the same symbol, and the state parameters between different clusters are different; in the time sequence, if the symbols of two adjacent state parameters are different, it can be determined that they belong to different sequence segments, thereby obtaining all sequence segments.

[0073] It should be noted that the first-order difference and the hierarchical clustering are both well-known prior art to those skilled in the art, and will not be described here; in other examples, the implementer can also use the slope to evaluate the trend or the sliding window-based method to calculate the variance of the state parameters in each sliding window, when the variance in the sliding window is the smallest, it means that it has a similar fluctuation trend, then it is determined as a sequence segment, and then all sequence segments are obtained; all of them are common technical means, and will not be described here.

[0074] Then the sequence length of the last sequence segment in time sequence, which is the nearest to the current time, is taken as the denominator, the total number of sequence segments is taken as the numerator, and the fractional ratio is taken as the fluctuation evaluation parameter of the corresponding index of the corresponding tower; then the fluctuation evaluation parameters of all indexes of each tower are accumulated to obtain the overall fluctuation parameter of each tower.

[0075] It should be noted that the total number of sequence segments and the sequence length of the last sequence segment in time sequence cannot be 0, so the fractional result is not 0 and always meaningful; in other examples, the implementer can also use other basic mathematical operations or related mapping means to combine the sequence length and the total number of sequence segments, which will not be described here.

[0076] In step S202, the tower sequence in the power transmission line project is obtained based on the connection of the towers in the power transmission line project; the fluctuation sequence is constructed according to the overall fluctuation parameters of each tower, the sequence order of the fluctuation sequence is the same as that of the tower sequence; and the tower sequence is divided according to the change of the fluctuation sequence, and all tower subsequences are obtained.

[0077] Since the towers in the power transmission line are usually arranged in sequence and then connected by the power transmission line, in an embodiment of the present application, the first tower at any end of the power transmission line is taken as the starting point, each tower is sequentially sorted to obtain the corresponding number, and thus the tower sequence is obtained; then the overall fluctuation parameter of each tower is taken as the sequence element, and the fluctuation sequence is constructed according to the order of the tower sequence; further, based on the change of the fluctuation sequence, the towers with similar state changes and close geographical positions are divided into a section, and all tower subsequences are obtained, so as to subsequently evaluate the state deviation of each tower relative to other towers.

[0078] In a preferred embodiment of the present application, the method for obtaining the tower subsequence comprises:

[0079] The first-order difference sequence of the fluctuation sequence is obtained, if the adjacent two sequence elements in the first-order difference sequence are of opposite signs, then the maximum number in the sequence numbers corresponding to the adjacent two sequence elements is taken as the cutting number;

[0080] The sequence element corresponding to the cutting number in the tower sequence is taken as the sequence right end point of each tower subsequence, and the adjacent next sequence element of each sequence right end point is taken as the sequence left end point of the next tower subsequence;

[0081] The tower sequence sub-section is obtained based on the sequence right end point and the sequence left end point of each tower subsequence.

[0082] To help understand the process of obtaining the tower subsequence, as an example, it is assumed that the tower sequence is {A, B, C, D, E, F, G}, and the corresponding fluctuation sequence is {1, 2, 3, 2, 4, 5, 4}, then the first-order difference sequence of the fluctuation sequence is {1, 1, -1, 2, 1, -1}, and all the obtained tower subsequences are {A, B, C}, {D, E, F}, and {G}.

[0083] In step S203, the state fluctuation coefficient of each tower is obtained according to the overall fluctuation parameter of each tower and the deviation degree of the overall fluctuation parameter of each tower relative to all the towers in the tower subsequence to which the tower belongs; the deviation degree is positively correlated with the state fluctuation coefficient.

[0084] If the overall fluctuation parameter of each tower reflects its state change, and if the state changes of all towers in the adjacent section that is relatively close and relatively similar are still relatively deviated, that is, there is a certain deviation from the state changes of all towers in the sequence of the tower, it means that the state fluctuation coefficient of the tower is more likely to deviate abnormally, and the state fluctuation coefficient is larger.

[0085] As an example, the calculation formula of the state fluctuation coefficient is: ; wherein i is the sequence number of the index; a is the sequence number of the tower; is the fluctuation state fluctuation coefficient of the a-th tower; is the total number of indexes; is the fluctuation evaluation parameter of the i-th index of the a-th tower; is the overall fluctuation parameter of the a-th tower; c is the sequence number of the tower in the tower sub-sequence to which the a-th tower belongs, and c can be a; is the total number of towers in the tower sub-sequence to which the a-th tower belongs; is the absolute value symbol.

[0086] In the calculation formula of the state fluctuation coefficient, reflects the average level of the overall fluctuation parameters of all towers in the tower sub-sequence to which the a-th tower belongs, and provides a reference basis for evaluating the deviation of the overall fluctuation parameter of each tower; the numerator reflects the deviation degree of the overall fluctuation parameter of each tower from the average level, and the smaller the average level corresponding to the denominator, the more stable the state of the adjacent tower, and the larger the deviation degree, the larger the state fluctuation coefficient of the tower.

[0087] It should be noted that since the fluctuation evaluation parameter of each tower under each index is not 0, the overall fluctuation parameter of each tower is not 0, and the fraction is always meaningful when it is used as the denominator.

[0088] Step S3, according to the fluctuation change of the state parameter in the time sequence under each index of each tower, combining the state fluctuation coefficient of each tower and the index importance of each index, the state influence weight of each index of each tower is obtained; at the current moment, according to the state influence weight of each index of each tower and the deviation of the state parameter of each tower under the same index from the state parameter of all towers, the final state influence weight of each index of each tower is obtained.

[0089] The more intense the fluctuation of the state parameter in the time sequence of each index of each tower is, the greater the state fluctuation coefficient of each tower and the index importance of each index are, which indicates that the influence on the change of the stable state of the tower is greater; therefore, the embodiment of the present application obtains the state influence weight of each index of each tower based on this, and the greater the state influence weight is, the more attention should be paid to the change of the state parameter of the tower under the index, that is, the attention weight of the corresponding index of the corresponding tower is higher, so as to capture the state change of the tower in time, thereby improving the state prediction accuracy of the tower.

[0090] Preferably, in an embodiment of the present application, the method for obtaining the state influence weight comprises:

[0091] The state fluctuation coefficient of each tower is multiplied by the index importance of each index, and the product is taken as a first weight parameter; the fluctuation evaluation parameter of each index of each tower is normalized, and the normalized result is taken as a second weight parameter; and the product of the first weight parameter and the second weight parameter is taken as the state influence weight of the corresponding index of the corresponding tower.

[0092] As an example, the calculation formula of the state influence weight is: ; wherein i is the serial number of the index; a is the serial number of the tower; is the state influence weight of the i-th index of the a-th tower; is the index importance of the i-th index; is the fluctuation state fluctuation coefficient of the a-th tower; is the fluctuation evaluation parameter of the i-th index of the a-th tower; norm() is a linear normalization function; is the first weight parameter; is the second weight parameter.

[0093] In this example, the parameter reflecting the state fluctuation of each tower and the importance of the index are multiplied and combined to obtain the state influence weight; the more important the index is, the greater the state influence weight is, and the higher the subsequent attention is; the more intense the state fluctuation of the tower is, the greater the state influence weight is, and the higher the subsequent attention is; in other examples, the implementer can also use basic mathematical operations such as addition or weighted summation or related mapping means to combine the three, and other normalization means can also be used, which are not described here.

[0094] In addition, the state parameters of each tower at the current moment are also crucial to subsequent state prediction, and the state changes of adjacent time sequences of the towers also have certain similarity. Therefore, the towers with relatively severe fluctuations at the current moment should be paid more attention to. Therefore, in the current moment, the final state influence weight of each index of each tower is obtained based on the state influence weight of each index of each tower and in combination with analysis of deviation of the state parameters of each tower under the same index from the state parameters of all towers. The final state influence weight combines the state deviation degree of the tower at the current moment and more accurately evaluates the attention degree of the state parameters of each index of each tower, so as to improve the state prediction accuracy.

[0095] Preferably, in an embodiment of the present application, the method for obtaining the final state influence weight comprises:

[0096] Under the same index, the fluctuation weight of each index of each tower is obtained according to the deviation of the state parameters of each tower from the average level of the state parameters of all towers. The corresponding state influence weight is weighted by using the fluctuation weight, and the weighted result is taken as the final state influence weight of each index of each tower.

[0097] As an example, the calculation formula of the final state influence weight is: ; wherein i is the serial number of the index; a is the serial number of the tower; is the final state influence weight of the i-th index of the a-th tower; is the state influence weight of the i-th index of the a-th tower; is the state parameter of the i-th index of the a-th tower at the current moment; is the average value of the state parameters of all towers under the i-th index at the current moment; is the absolute value symbol; is the fluctuation weight of the i-th index of the a-th tower; is a standard normalization function.

[0098] In the calculation formula of the final state influence weight, reflects the deviation of the state parameters of the a-th tower from the average level of the state parameters of all towers under the same index. The greater the deviation, the greater the fluctuation weight, and the greater the necessity of subsequent attention. Then, the ratio of the fluctuation weight to the state influence weight is multiplied to obtain the final state influence weight. In this example, the sum of the weighted results of all indexes of each tower is taken as the denominator, and the weighted result under each index is taken as the numerator, and the ratio is normalized to make the sum of the final state influence weights of all indexes of each tower equal to 1.

[0099] At this point, the final state influence weight of each tower under each index at the current time is obtained, that is, the attention degree of each index of each tower to the state parameter, so as to improve the accuracy of subsequent state prediction.

[0100] Step S4, according to the final state influence weight and the current state vector, the predicted state vector of each tower at the next collection time is obtained.

[0101] After obtaining the final state influence weight of each tower under each index and the current state vector of each tower at the current time, the state prediction of each tower at the next collection time can be performed.

[0102] Preferably, in an embodiment of the present application, considering that the Kalman filter is a recursive algorithm for estimating the state of a dynamic system, which can be used for prediction and state estimation of time series data, the current state vector of the corresponding tower can be weighted by using the final state influence weight, the attention degree to each index is adjusted, and then the Kalman filter can more accurately capture the state change of the system, thereby improving the accuracy of the output prediction result; based on this, the method for obtaining the predicted state vector comprises:

[0103] The current state vector of the corresponding tower is weighted by using the final state influence weight, and the weighted result is taken as the input of the Kalman filter, and the predicted state vector of the corresponding tower at the next collection time is output.

[0104] It should be noted that the acquisition of the weighted result and the application of the Kalman filter are well known to those skilled in the art, and will not be described here.

[0105] At this point, the predicted state vector of each tower at the next collection time is obtained, and the implementer can evaluate the safety risk of the power transmission line project through the predicted state vector, such as whether the load exceeds the preset upper limit of the load, so as to take optimization measures to reduce the safety hidden danger.

[0106] In summary, the present application first obtains the time sequence of the state parameter of each tower under each index and the current state vector of each tower; then, according to the change correlation between different indexes of different towers, the index importance of each index is obtained; the state fluctuation coefficient of each tower is further analyzed to obtain the state influence weight of each index of each tower; then, at the current time, the final state influence weight of each tower under each index is obtained, so as to obtain the predicted state vector of each tower at the next collection time in combination with the current state vector. The present application evaluates the attention weight of each index of each tower by analyzing the change correlation and historical fluctuation of the state parameter under the multi-dimensional index of the tower, so as to adjust the input of the Kalman filter algorithm, so that the algorithm captures the index characteristics that have a great influence on the state change of the tower, and then improves the prediction accuracy of the state of each tower in the power transmission line project.

[0107] It should be noted that the progressive order of the above-mentioned embodiments of the application is only for the purpose of description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0108] Each of the embodiments in the specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment mainly explains the difference from other embodiments.

Claims

1. A method for predicting mutual influence of power transmission line engineering indexes based on multi-dimensional data analysis, characterized in that, The method comprises: At the current time, obtaining the time sequence of the state parameters of each tower in each index in the power transmission line project in a preset historical period, and the current state vector of each tower; the current state vector comprises the state parameters in each index; According to the change correlation between all towers in different indexes, the index importance of each index is obtained, and the index importance is the fractional ratio of the difference between the first parameter and the second parameter as the numerator and the difference between the third parameter and the fourth parameter as the denominator; the first parameter is the sum value of the product of the first dependent reference value and the second dependent reference value of all towers, the second parameter is the product of the sum value of the first dependent reference value of all towers and the sum value of the second dependent reference value of all towers, the third parameter is the square sum of the second dependent reference value of all towers, and the fourth parameter is the square sum of the second dependent reference value of all towers; the first dependent reference value is the sum value of all elements in each row of the state transition matrix, and the second dependent reference value is the sum value of all elements in each column of the state transition matrix; each row and each column of the state transition matrix corresponds to an index, and the state transition matrix is obtained based on the Kalman filter and the state vectors at all adjacent collection times; According to the fluctuation change of the state parameters in the time sequence of each tower in each index, the fluctuation evaluation parameter of each index of each tower is obtained; the overall fluctuation parameter of each tower is obtained by comprehensively considering the fluctuation evaluation parameters of all indexes of each tower; Based on the connection of the towers in the power transmission line project, the tower sorting sequence in the power transmission line project is obtained; the fluctuation sequence is constructed according to the overall fluctuation parameter of each tower, and the sequence order of the fluctuation sequence is the same as that of the tower sorting sequence; according to the change of the fluctuation sequence, the tower sorting sequence is divided to obtain all tower subsequences; According to the overall fluctuation parameter of each tower, the state fluctuation coefficient of each tower is obtained relative to the deviation degree of the overall fluctuation parameter of all towers in the tower subsequence to which the tower belongs; the deviation degree is positively correlated with the state fluctuation coefficient; According to the fluctuation change of the time sequence of each tower in each index, combined with the state fluctuation coefficient of each tower and the index importance of each index, the state influence weight of each index of each tower is obtained; at the current time, according to the state influence weight of each index of each tower and the deviation of the state parameters of each tower in the same index relative to the state parameters of all towers, the final state influence weight of each index of each tower is obtained; The final state influence weight is used to weight the current state vector of the corresponding tower, and the weighting result is taken as the input of the Kalman filter, and the predicted state vector of the corresponding tower at the next collection time is output. The method for obtaining the tower sub-sequences comprises: obtaining a first-order difference sequence of the fluctuation sequence; if adjacent two sequence elements in the first-order difference sequence are of different signs, taking the maximum sequence number of the adjacent two sequence elements as a cutting sequence number; taking a sequence element corresponding to the cutting sequence number in the tower sequence as a right end point of each tower sub-sequence; taking a next adjacent sequence element of each right end point as a left end point of a next tower sub-sequence; and obtaining a tower sequence sub-section based on the right end point and the left end point of each tower sub-sequence. The method for obtaining the final state influence weight comprises: under the same index, obtaining a fluctuation weight of each index of each tower according to a deviation of a state parameter of each tower from an average level of state parameters of all towers; and weighting a corresponding state influence weight by using the fluctuation weight, and taking a weighting result as a final state influence weight of each index of each tower.

2. The method of claim 1, wherein the method is based on multi-dimensional data analysis of power transmission line engineering indicators. The method for obtaining the fluctuation evaluation parameter comprises: segmenting a time sequence of each tower under each index according to local fluctuation similarity of the time sequence to obtain all sequence sub-sections; obtaining a fluctuation evaluation parameter of each index of each tower according to a total number of the sequence sub-sections and a sequence length of a last sequence sub-section in time sequence; the sequence length is negatively correlated with the fluctuation evaluation parameter, and the total number is positively correlated with the fluctuation evaluation parameter.

3. The method of claim 1, wherein the method is characterized by: The method for obtaining the state influence weight comprises: multiplying a state fluctuation coefficient of each tower and an index importance of each index to obtain a first weight parameter; normalizing the fluctuation evaluation parameter of each index of each tower to obtain a second weight parameter; and multiplying the first weight parameter and the second weight parameter to obtain a state influence weight of a corresponding index of a corresponding tower.

4. The method of claim 1, wherein the method is based on multi-dimensional data analysis of power transmission line engineering indicators. The index at least comprises a load, a tension ratio, a height, a span, a weight and a dancing amplitude.

Images