OPGW optical cable life prediction method and system

Abstract

The application provides an OPGW optical cable life prediction method and system, and relates to the field of data processing.The method comprises the following steps: determining a plurality of degradation performance factor groups based on a multi-dimensional performance degradation performance factor set; for each degradation performance factor group, establishing a residual life prediction model corresponding to the degradation performance factor group; collecting multi-dimensional performance data of the OPGW optical cable based on the multi-dimensional performance degradation performance factor set; for each degradation performance factor group, extracting a performance data set corresponding to the degradation performance factor group from the multi-dimensional performance data of the OPGW optical cable, decomposing the performance data set corresponding to the degradation performance factor group by using a variational mode decomposition algorithm to obtain a decomposition result, and predicting the residual life of a reference OPGW optical cable based on the decomposition result by using the residual life prediction model corresponding to the degradation performance factor group; and determining the residual life of the OPGW optical cable based on the residual life of the reference OPGW optical cable corresponding to each degradation performance factor group, which has the advantage of improving the accuracy of OPGW optical cable life prediction.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of data processing, in particular to an OPGW optical cable life prediction method and system. BACKGROUND

[0002] OPGW optical cable (Optical Fiber Composite Overhead Ground Wire) is a kind of special optical cable combining optical fiber communication unit and overhead ground wire, mainly used in high-voltage transmission lines, with double functions of communication transmission and lightning protection grounding. When working, OPGW optical cable is often in complex and harsh environment, and bears the influence of disasters such as icing, lightning, strong wind, rainstorm and high temperature, and becomes a weak link in the power transmission network. Whether the health of OPGW optical cable operation state is directly related to the communication safety of power system, with the increasing of ultra-high voltage, long distance and important crossing overhead transmission lines, the safety problem of OPGW optical cable in complex environment is increasingly prominent. The life of OPGW optical cable is related to the material, structure, line condition, natural environment, construction method and whether lightning strike of the optical cable. The life of OPGW optical cable running more than 15 years is quite different, some of which have been retired after lightning breakage, some of which are still in good condition and can continue to run, and some of which are in poor condition and need to be evaluated for life. If a one-size-fits-all approach is taken to OPGW optical cable maintenance or retirement, it will cause huge economic losses. Therefore, life evaluation of OPGW optical cable helps to improve the fine management of optical cable and provides corresponding technical support for the whole life cycle management of OPGW optical cable.

[0003] In the prior art, the line dynamic bending stress is calculated based on the energy balance model of OPGW optical cable, and the line life is evaluated in combination with the Miner cumulative damage theory. However, many ideal conditions need to be set in the process of establishing the energy balance model of OPGW optical cable, which leads to deviation in calculation of line wind vibration frequency and amplitude, and inaccuracy in calculation of dynamic bending stress, affecting the precision of life evaluation.

[0004] Therefore, it is necessary to provide an OPGW optical cable life prediction method and system for improving the accuracy of OPGW optical cable life prediction. SUMMARY

[0005] The application provides an OPGW optical cable life prediction method, comprising: determining a multi-dimensional performance degradation performance factor set, wherein the multi-dimensional performance degradation performance factor set comprises a plurality of mechanical degradation performance factors, electrical degradation performance factors and transmission degradation performance factors; determining a plurality of degradation performance factor groups based on the multi-dimensional performance degradation performance factor set; for each degradation performance factor group, establishing a residual life prediction model corresponding to the degradation performance factor group; based on the multi-dimensional performance degradation performance factor set, collecting multi-dimensional performance data of the OPGW optical cable; for each degradation performance factor group, extracting a performance data set corresponding to the degradation performance factor group from the multi-dimensional performance data of the OPGW optical cable, decomposing the performance data set corresponding to the degradation performance factor group by a variational mode decomposition algorithm to obtain a decomposition result, and predicting the residual life of the reference OPGW optical cable based on the decomposition result by the residual life prediction model corresponding to the degradation performance factor group; and determining the residual life of the OPGW optical cable based on the residual life of the reference OPGW optical cable corresponding to each degradation performance factor group.

[0006] Further, the multi-dimensional performance degradation performance factor set is determined, comprising: determining a plurality of to-be-screened degradation performance factors; obtaining performance test data of a plurality of sample OPGW optical cables based on the plurality of to-be-screened degradation performance factors, wherein the residual lives of the plurality of sample OPGW optical cables are different; determining a correlation coefficient of each to-be-screened degradation performance factor and the residual life based on the performance test data of the plurality of sample OPGW optical cables; screening a plurality of effective degradation performance factors from the plurality of to-be-screened degradation performance factors based on the correlation coefficient of each to-be-screened degradation performance factor and the residual life; determining a correlation coefficient of any two effective degradation performance factors based on the performance test data of the plurality of sample OPGW optical cables; for each effective degradation performance factor, calculating a redundancy value of the effective degradation performance factor based on the correlation coefficient of the effective degradation performance factor and any other effective degradation performance factor; clustering the plurality of effective degradation performance factors according to the correlation coefficient of any two effective degradation performance factors to generate a plurality of factor clusters; and determining the multi-dimensional performance degradation performance factor set according to the redundancy value of each effective degradation performance factor and the plurality of factor clusters, wherein the multi-dimensional performance degradation performance factor set comprises a plurality of degradation performance factors.

[0007] Further, the set of multi-dimensional performance degradation factors is determined according to the redundancy values of each valid degradation performance factor and the plurality of factor clusters, including: establishing factor sampling constraints and cluster coverage constraints; establishing the factor sampling constraints and the cluster coverage constraints, sampling the plurality of factor clusters to generate a primary population; establishing a fitness function, wherein the fitness function is related to the redundancy values of the valid degradation performance factors included in an individual and the correlation coefficients of any two valid degradation performance factors; for each individual, determining the coverage of the plurality of factor clusters by the individual, calculating the fitness value of the individual based on the fitness function; determining a plurality of retained individuals based on the coverage of the plurality of factor clusters by each individual and the fitness value; for each retained individual, determining a factor cluster coverage vector of the retained individual; calculating the cosine similarity of the factor cluster coverage vectors of any two retained individuals; performing crossover on the plurality of retained individuals according to the cosine similarity of the factor cluster coverage vectors of any two retained individuals; performing mutation on the plurality of retained individuals according to the coverage of the plurality of factor clusters by each retained individual; updating the population according to the results of the crossover and the mutation, and iteratively optimizing until a first end condition is met.

[0008] Further, based on the set of multi-dimensional performance degradation factors, a plurality of degradation performance factor groups is determined, including: based on the correlation coefficients of any two degradation performance factors, determining the plurality of degradation performance factor groups by a nested clustering algorithm.

[0009] Further, based on the set of multi-dimensional performance degradation factors, multi-dimensional performance data of the OPGW optical cable is collected, including at least: obtaining the attenuation coefficient of the OPGW optical cable by using optical time domain reflection technology; obtaining the polarization state of the OPGW optical cable by using polarization optical time domain reflection technology; measuring the metal conductor resistance of the OPGW optical cable by using a four-terminal method; measuring the dispersion parameter of the OPGW optical cable by using an optical frequency domain reflectometer; and measuring the grounding resistance value of the OPGW optical cable by using a grounding resistance tester.

[0010] Further, the performance data set corresponding to the degradation performance factor group is decomposed by a variational mode decomposition algorithm to obtain a decomposition result, including: determining the key intrinsic mode function serial number of each degradation performance factor included in the degradation performance factor group based on the performance test data of the plurality of sample OPGW optical cables and the variational mode decomposition algorithm; decomposing the performance data set corresponding to the degradation performance factor group by the variational mode decomposition algorithm to obtain a plurality of intrinsic mode functions of each degradation performance factor included in the degradation performance factor group; and extracting the key intrinsic mode function of each degradation performance factor included in the degradation performance factor group from the plurality of intrinsic mode functions of each degradation performance factor included in the degradation performance factor group based on the key intrinsic mode function serial number of each degradation performance factor included in the degradation performance factor group.

[0011] Further, based on the performance test data of the plurality of sample OPGW optical cables and the variational mode decomposition algorithm, determining the key intrinsic mode function sequence number of each degradation performance factor included in the degradation performance factor group comprises: based on the performance test data of the plurality of sample OPGW optical cables and the variational mode decomposition algorithm, determining the correlation coefficient of each intrinsic mode function sequence number of each degradation performance factor and the remaining life and the correlation coefficient of the intrinsic mode function sequence numbers of any two degradation performance factors; based on the correlation coefficient of each intrinsic mode function sequence number of each degradation performance factor and the remaining life and the correlation coefficient of the intrinsic mode function sequence numbers of any two degradation performance factors, determining the key intrinsic mode function sequence number of each degradation performance factor included in the degradation performance factor group.

[0012] Further, the remaining life prediction model comprises: an input layer for inputting the key intrinsic mode function of each degradation performance factor included in the degradation performance factor group; a convolution layer comprising a plurality of convolution blocks for extracting local time-frequency features of the key intrinsic mode function of each degradation performance factor; a feature fusion layer for fusing the local time-frequency features of the key intrinsic mode function of each degradation performance factor to construct a global degradation feature vector; a fully connected layer for mapping the global degradation feature vector to the remaining life of the OPGW optical cable through nonlinear transformation and stepwise dimension reduction; and an output layer for outputting the remaining life of the OPGW optical cable.

[0013] Further, based on the remaining life of the reference OPGW optical cable corresponding to each degradation performance factor group, determining the remaining life of the OPGW optical cable comprises: based on the performance test data of the plurality of sample OPGW optical cables, establishing a polynomial multiple regression model, wherein the independent variable of the polynomial multiple regression model is the remaining life of the reference OPGW optical cable corresponding to each degradation performance factor group, and the dependent variable of the polynomial multiple regression model is the remaining life of the OPGW optical cable; and based on the remaining life of the reference OPGW optical cable corresponding to each degradation performance factor group, determining the remaining life of the OPGW optical cable through the polynomial multiple regression model.

[0014] The application provides an OPGW optical cable life prediction system for the OPGW optical cable life prediction method, comprising: a factor determination module configured to determine a multi-dimensional performance degradation performance factor set, wherein the multi-dimensional performance degradation performance factor set comprises a plurality of mechanical degradation performance factors, electrical degradation performance factors and transmission degradation performance factors; a factor grouping module configured to determine a plurality of degradation performance factor groups based on the multi-dimensional performance degradation performance factor set; a data acquisition module configured to acquire multi-dimensional performance data of the OPGW optical cable based on the multi-dimensional performance degradation performance factor set; and a life prediction module configured to, for each degradation performance factor group, establish a residual life prediction model corresponding to the degradation performance factor group, extract a performance data set corresponding to the degradation performance factor group from the multi-dimensional performance data of the OPGW optical cable, decompose the performance data set corresponding to the degradation performance factor group by using a variational mode decomposition algorithm to obtain a decomposition result, predict a residual life of a reference OPGW optical cable based on the decomposition result by using the residual life prediction model corresponding to the degradation performance factor group, and determine the residual life of the OPGW optical cable based on the residual life of the reference OPGW optical cable corresponding to each degradation performance factor group.

[0015] Compared with the prior art, the OPGW optical cable life prediction method and system provided by the application has at least the following beneficial effects:

[0016] The multi-dimensional performance degradation performance factor set (covering mechanical, electrical and transmission factors) comprehensively reflects the complex degradation process of the OPGW optical cable, avoiding the one-sidedness of single-factor prediction. The variational mode decomposition algorithm is used to decompose the performance data set, effectively extract degradation features, reduce noise interference, and improve the accuracy of residual life prediction. Independent residual life prediction models are established for different degradation performance factor groups, and the prediction results of multiple groups are fused to enhance the robustness of the model. The prediction results of each factor group are integrated by using a polynomial multiple regression model to capture the nonlinear interaction between the degradation factors, which is more in line with the actual degradation law. BRIEF DESCRIPTION OF DRAWINGS

[0017] The present specification will be further illustrated in the form of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. These embodiments are not limiting, and in these embodiments, the same reference numbers represent the same structures, wherein:

[0018] Figure 1 is a flowchart of an OPGW optical cable life prediction method according to some embodiments of the present specification;

[0019] Figure 2 is a flowchart of determining a multi-dimensional performance degradation performance factor set according to some embodiments of the present specification;

[0020] Figure 3is a flowchart of a process of determining a set of multi-dimensional performance degradation performance factors according to some embodiments of the present specification;

[0021] Figure 4 is a structural diagram of a remaining life prediction model according to some embodiments of the present specification;

[0022] Figure 5 is an effect diagram of a remaining life prediction model according to some embodiments of the present specification;

[0023] Figure 6 is an effect diagram of a polynomial multiple regression model according to some embodiments of the present specification;

[0024] Figure 7 is a module diagram of an OPGW optical cable life prediction system according to some embodiments of the present specification. DETAILED DESCRIPTION

[0025] In order to more clearly illustrate the technical solutions of the embodiments of the present specification, the following will briefly introduce the drawings needed to be used in the embodiment description. Obviously, the drawings in the following description are only some examples or embodiments of the present specification, and for those skilled in the art, the present specification can be applied to other similar scenarios without creative labor. Unless it is obvious from the language environment or otherwise stated, the same reference numbers in the drawings represent the same structure or operation.

[0026] Figure 1 is a flowchart of an OPGW optical cable life prediction method according to some embodiments of the present specification, as shown in Figure 1 The OPGW optical cable life prediction method can include the following steps:

[0027] Step 1, determining a set of multi-dimensional performance degradation performance factors.

[0028] The set of multi-dimensional performance degradation performance factors includes a plurality of mechanical degradation performance factors, electrical degradation performance factors and transmission degradation performance factors.

[0029] Mechanical degradation is a structural damage of OPGW optical cable caused by external force (such as wind vibration, icing, construction damage), which directly affects the physical integrity and mechanical strength of the optical cable. As an example, the mechanical degradation performance factor can include a tensile strain factor, which represents the degree of deformation of the optical cable under tension, expressed in percentage. Under normal working tension (such as 40% RTS), the excess length of the optical fiber can completely absorb the elongation of the cable body, and the optical fiber strain is close to zero or maintained at a very low level (≤0.05%). After long-term operation, the initial elongation, creep or environmental load causes the excess length to decrease, and the optical fiber begins to bear part of the strain. For example, the optical fiber strain can increase to 0.12%-0.19%. When the optical fiber excess length is completely exhausted, the elongation of the cable body is directly transmitted to the optical fiber, causing the strain to rise sharply. For example, the optical fiber strain can be ≥0.19%.

[0030] The electrical degradation performance factor refers to the quantifiable or observable indicators generated by the degradation of the electrical performance of the OPGW optical cable during long-term operation, which directly reflects the degree of degradation of the optical cable in the electrical aspect. As an example, the electrical degradation performance factor can include a sheath resistance factor, a grounding resistance factor, an electromagnetic shielding performance factor, etc. Among them, the sheath material (such as aluminum-clad steel or aluminum alloy) causes the sheath resistance to increase due to electrochemical corrosion, environmental erosion or poor contact. The grounding resistance factor refers to the resistance value between the OPGW optical cable and the ground, reflecting the reliability and effectiveness of the optical cable grounding system. The electromagnetic shielding performance factor refers to the attenuation ability of the OPGW optical cable to external electromagnetic interference, reflecting the shielding effect of its sheath and metal structure.

[0031] The transmission degradation performance factor is a quantifiable or observable indicator reflecting the degradation of the transmission performance of the OPGW optical cable due to various factors during long-term operation. As an example, the transmission degradation performance factor can include a fiber attenuation factor, a splicing loss factor, a polarization state rotation speed factor, etc. Among them, fiber attenuation refers to the loss of optical power due to scattering, absorption, etc. during the propagation of light in the optical fiber. The splicing loss factor refers to the loss of optical power caused by end face contamination, improper splicing parameter setting, etc. during the splicing process of the optical fiber. Experiments show that under the action of Karman vortex vibration (wind vibration), the polarization state rotation speed of a healthy OPGW optical cable can reach 794.2 rad / s, while an early degradation OPGW optical cable may cause the rotation speed fluctuation range to expand due to birefringence changes, and even appear short-term abnormal acceleration (such as locally reaching 800-900 rad / s). A medium-term degradation optical cable may cause the rotation speed to break through the design limit (such as reaching 8000-9000 rad / s) due to stress concentration, and even cause polarization mismatch, optical fiber breakage or complete failure of the sheath, resulting in interruption of the optical signal transmission path, and the polarization state cannot be maintained. Near the breaking point, the polarization state may present random jumps, and the rotation speed cannot be measured.

[0032] Figure 2is a flowchart of a process for determining a set of multi-dimensional performance degradation performance factors according to some embodiments of the present specification, as shown in Figure 2 As preferred, step 1 specifically includes:

[0033] The plurality of to-be-screened degradation performance factors can be determined in combination with historical operation data, industry standards and expert experience. For example, the plurality of to-be-screened degradation performance factors can include sheath tensile strength, fitting fatigue crack, fiber excess length, sheath resistance, grounding resistance, electromagnetic shielding effectiveness, discharge intensity, fiber attenuation, splicing loss, dispersion coefficient, polarization mode dispersion, etc.

[0034] Based on the plurality of to-be-screened degradation performance factors, performance test data of a plurality of sample OPGW optical cables is obtained, wherein the plurality of sample OPGW optical cables have different residual lifespans, and wherein the performance test data of the sample OPGW optical cables can include values of each to-be-screened degradation performance factor corresponding to the sample OPGW optical cables.

[0035] Based on the performance test data of the plurality of sample OPGW optical cables, a correlation coefficient of each to-be-screened degradation performance factor and the residual lifespan is determined. Specifically, for each to-be-screened degradation performance factor, the value of the to-be-screened degradation performance factor in each sample OPGW optical cable and the residual lifespan of each sample OPGW optical cable are substituted into a correlation coefficient calculation formula (e.g., Pearson correlation coefficient, Spearman rank correlation coefficient, etc.) to calculate the correlation coefficient of the to-be-screened degradation performance factor and the residual lifespan.

[0036] Based on the correlation coefficient of each to-be-screened degradation performance factor and the residual lifespan, a plurality of effective degradation performance factors is screened from the plurality of to-be-screened degradation performance factors. For example, the to-be-screened degradation performance factor with a correlation coefficient greater than a first threshold value (e.g., 0.3) is regarded as an effective degradation performance factor, wherein the first threshold value can be determined through a large amount of experimental data.

[0037] Based on the performance test data of the plurality of sample OPGW optical cables, a correlation coefficient of any two effective degradation performance factors is determined. Specifically, for any two effective degradation performance factors, the values of the two effective degradation performance factors in each sample OPGW optical cable are substituted into a correlation coefficient calculation formula (e.g., Pearson correlation coefficient, Spearman rank correlation coefficient, etc.) to calculate the correlation coefficient of the two effective degradation performance factors.

[0038] For each valid degradation performance factor, a redundancy value of the valid degradation performance factor is calculated based on a correlation coefficient between the valid degradation performance factor and any other valid degradation performance factor, specifically, a mean value of the correlation coefficient between the valid degradation performance factor and any other valid degradation performance factor is calculated as the redundancy value of the valid degradation performance factor;

[0039] According to the correlation coefficient between any two valid degradation performance factors, the plurality of valid degradation performance factors are clustered to generate a plurality of factor clusters, specifically, a clustering distance between any two valid degradation performance factors can be determined according to the correlation coefficient between the two valid degradation performance factors, the greater the correlation coefficient between the two valid degradation performance factors, the shorter the clustering distance between the two valid degradation performance factors, and the plurality of valid degradation performance factors are clustered to generate a plurality of factor clusters according to the clustering distance between any two valid degradation performance factors by a clustering algorithm (for example, K-Means clustering, hierarchical clustering, etc.);

[0040] According to the redundancy value of each valid degradation performance factor and the plurality of factor clusters, a multi-dimensional performance degradation performance factor set is determined, wherein the multi-dimensional performance degradation performance factor set includes a plurality of degradation performance factors.

[0041] Figure 3 is a flowchart of determining a multi-dimensional performance degradation performance factor set based on an improved genetic algorithm according to some embodiments of the present specification, as shown in Figure 3 According to the redundancy value of each valid degradation performance factor and the plurality of factor clusters, a multi-dimensional performance degradation performance factor set is determined, including:

[0042] Factor sampling constraints and cluster coverage constraints are established, wherein the factor sampling constraints can represent the maximum number of valid degradation performance factors extracted from each factor cluster, for example, 2, 3, etc., and the cluster coverage constraints can represent the minimum proportion of valid degradation performance factors included in each individual covering a plurality of factor clusters, for example, 70%, 80%, only as an example, there are factor cluster A1, factor cluster A2, factor cluster A3, and factor cluster A4, and the valid degradation performance factors included in the individual are from factor cluster A1 and factor cluster A2, respectively, so the cluster coverage of the individual is 50%, which does not meet the requirement of the cluster coverage constraint of 70%;

[0043] Factor sampling constraints and cluster coverage constraints are established, and a plurality of factor clusters are sampled to generate a primary population, for example, according to the cluster coverage constraint, k factor clusters are randomly selected from M factor clusters, for each selected factor cluster, at least one valid degradation performance factor is sampled from the factor cluster according to the factor sampling constraint to generate an individual, and the step is repeated until the generation of the primary population is completed.

[0044] establishing a fitness function, wherein the fitness function is related to redundancy values of the valid degradation performance factors included in the individual and correlation coefficients of any two valid degradation performance factors, for example, the greater the mean of the redundancy values of the valid degradation performance factors included in the individual and the greater the correlation coefficients of any two valid degradation performance factors, the smaller the value of the fitness function;

[0045] for each individual, determining coverage of the individual to the plurality of factor clusters, and calculating a fitness value of the individual based on the fitness function;

[0046] based on the coverage of each individual to the plurality of factor clusters and the fitness value, determining a plurality of reserved individuals, specifically, the greater the coverage of the individual to the plurality of factor clusters and the higher the fitness value, the greater the probability of being a reserved individual;

[0047] for each reserved individual, determining a factor cluster coverage vector of the reserved individual, wherein the factor cluster coverage vector can be composed of coverage identifiers of each factor cluster, if the individual covers a factor cluster, the coverage identifier of the factor cluster is 1, otherwise it is 0;

[0048] calculating the cosine similarity of the factor cluster coverage vectors of any two reserved individuals;

[0049] crossing the plurality of reserved individuals according to the cosine similarity of the factor cluster coverage vectors of any two reserved individuals, specifically, the smaller the cosine similarity of two reserved individuals, the higher the probability of crossing;

[0050] mutating the plurality of reserved individuals according to the coverage of each reserved individual to the plurality of factor clusters, specifically, the higher the coverage of an individual, the higher the probability of mutation;

[0051] updating the population according to the crossing and mutation results, and performing iterative optimization until a first end condition is met, wherein the first end condition can be that the number of iterations reaches a maximum number of iterations, fitness converges, etc., and after the end, including the plurality of valid degradation performance factors in the individual with the highest fitness as the set of multi-dimensional performance degradation performance factors.

[0052] The above process enforces individual coverage of multiple factor clusters by cluster coverage constraints (e.g., requiring coverage ≥ 70%), avoiding the problem of excessive concentration of effective degradation performance factors in a small number of clusters. Ensuring that the factor set can reflect the multi-dimensional characteristics of OPGW cable degradation, improving the reliability of residual life prediction. In the fitness function, high redundancy values are penalized (e.g., the greater the average redundancy, the lower the fitness), eliminating redundant information, and suppressing highly correlated factor pairs through the fitness function (e.g., the greater the correlation coefficient, the lower the fitness), reducing information overlap. Generate a more compact factor set, reduce computational complexity, while retaining key degradation performance information. Prioritize individuals with large differences in crossover coverage vectors (low cosine similarity), enhance population diversity, and avoid premature convergence. Increase the mutation probability for high coverage individuals to prevent local optimization, while exploring uncovered factor clusters, so that the final generated multi-dimensional performance degradation factor set has the characteristics of high coverage, low redundancy, and strong independence.

[0053] Step 2, based on the multi-dimensional performance degradation factor set, determine a plurality of degradation performance factor groups.

[0054] Specifically includes:

[0055] By genetic algorithm, based on the correlation coefficient of any two degradation performance factors, determine a plurality of degradation performance factor groups, wherein the fitness function of the genetic algorithm aims to minimize the average of the correlation coefficient of any two degradation performance factors included in each degradation performance factor group.

[0056] Step 3, for each degradation performance factor group, establish a residual life prediction model corresponding to the degradation performance factor group.

[0057] Step 4, based on the multi-dimensional performance degradation factor set, collect multi-dimensional performance data of the OPGW cable.

[0058] At least includes:

[0059] Using optical time domain reflectometry to obtain the attenuation coefficient of the OPGW cable;

[0060] Using polarization optical time domain reflectometry to obtain the polarization state of the OPGW cable;

[0061] Using four-terminal method to measure the metal wire resistance of the OPGW cable;

[0062] Using optical frequency domain reflectometer to measure the dispersion parameter of the OPGW cable;

[0063] Using ground resistance tester to measure the ground resistance value of the OPGW cable.

[0064] Other technical means can also be included, for example, the tensile strain of the OPGW optical cable is measured by the fiber grating sensor. The attenuation coefficient is calculated by directly measuring the difference in optical power at the input and output ends of the optical fiber. The attenuation is calculated by comparing the difference in readings of the optical power meter after directly connecting the light source and through the optical fiber. The splicing loss is measured by the bidirectional average splicing loss test method of the optical time domain reflectometer, etc.

[0065] The multi-dimensional performance data of the OPGW optical cable can include the value of each degradation performance factor included in the set of multi-dimensional performance degradation performance factors at multiple time points.

[0066] Step 5, for each degradation performance factor group, from the multi-dimensional performance data of the OPGW optical cable, extract the performance data set corresponding to the degradation performance factor group, decompose the performance data set corresponding to the degradation performance factor group by the variational mode decomposition algorithm to obtain the decomposition result, and predict the remaining life of the reference OPGW optical cable based on the decomposition result by the remaining life prediction model corresponding to the degradation performance factor group.

[0067] As preferred, the decomposition of the performance data set corresponding to the degradation performance factor group by the variational mode decomposition algorithm to obtain the decomposition result comprises:

[0068] Based on the performance test data of the plurality of sample OPGW optical cables and the variational mode decomposition algorithm, determine the key intrinsic mode function sequence number of each degradation performance factor included in the degradation performance factor group;

[0069] The decomposition of the performance data set corresponding to the degradation performance factor group by the variational mode decomposition algorithm obtains a plurality of intrinsic mode functions of each degradation performance factor included in the degradation performance factor group;

[0070] Based on the key intrinsic mode function sequence number of each degradation performance factor included in the degradation performance factor group, extract the key intrinsic mode function of each degradation performance factor included in the degradation performance factor group from the plurality of intrinsic mode functions of each degradation performance factor included in the degradation performance factor group.

[0071] As preferred, based on the performance test data of the plurality of sample OPGW optical cables and the variational mode decomposition algorithm, determining the key intrinsic mode function sequence number of each degradation performance factor included in the degradation performance factor group comprises:

[0072] Based on the performance test data of the plurality of sample OPGW optical cables and the variational mode decomposition algorithm, determine the correlation coefficient of each intrinsic mode function sequence number of each degradation performance factor and the remaining life and the correlation coefficient of the intrinsic mode function sequence number of any two degradation performance factors;

[0073] Based on the correlation coefficient of each eigenmode function serial number of each degradation performance factor and the remaining life and the correlation coefficient of the eigenmode function serial numbers of any two degradation performance factors, the key eigenmode function serial number of each degradation performance factor included in the degradation performance factor group is determined.

[0074] Specifically, for each sample OPGW optical cable, the residual time life of each sample OPGW optical cable is sorted from large to small to generate a sorting result, and based on the performance test data of the plurality of sample OPGW optical cables and the sorting result, a value sequence of each degradation performance factor corresponding to the sample OPGW optical cable is determined, wherein one element in the value sequence represents the value of the degradation performance factor corresponding to one sample OPGW optical cable, and the sorting order of the value of the degradation performance factor corresponding to the sample OPGW optical cable in the value sequence is consistent with the sorting result.

[0075] For each degradation performance factor, the value sequence of the degradation performance factor is subjected to variational mode decomposition to obtain a plurality of eigenmode functions corresponding to the degradation performance factor, and for each eigenmode function, the value of the eigenmode function corresponding to each sample OPGW optical cable and the residual life of each sample OPGW optical cable are substituted into a correlation coefficient calculation formula (such as Pearson correlation coefficient, Spearman rank correlation coefficient, etc.) to calculate the correlation coefficient of the eigenmode function serial number and the residual life.

[0076] For two degradation performance factors included in the degradation performance factor group, the value of the eigenmode function of one degradation performance factor corresponding to each sample OPGW optical cable and the value of the eigenmode function of another degradation performance factor corresponding to each sample OPGW optical cable are substituted into a correlation coefficient calculation formula (such as Pearson correlation coefficient, Spearman rank correlation coefficient, etc.) to calculate the correlation coefficient of the eigenmode function serial numbers of the two degradation performance factors. It can be understood that if one degradation performance factor has 4 eigenmode functions and another degradation performance factor has 4 eigenmode functions, the correlation coefficient of each eigenmode function of one degradation performance factor and each eigenmode function of another degradation performance factor needs to be calculated.

[0077] The key intrinsic mode function sequence number of each degradation performance factor included in the degradation performance factor group can be determined by a genetic algorithm based on the correlation coefficient of each intrinsic mode function sequence number of each degradation performance factor and the remaining useful life and the correlation coefficient of the intrinsic mode function sequence numbers of any two degradation performance factors, wherein the fitness function target of the genetic algorithm is to achieve a balance of the maximum sum of the correlation coefficients and the minimum mean of the correlation coefficients of the key intrinsic mode function sequence numbers of any two degradation performance factors. As an example, assuming that the total number of degradation performance factors is 3, which are degradation performance factor B1, degradation performance factor B2, and degradation performance factor B3, the key intrinsic mode function sequence number of the degradation performance factor B1 in the individual is IMF11, the key intrinsic mode function sequence number of the degradation performance factor B2 is IMF21, and the key intrinsic mode function sequence number of the degradation performance factor B3 is IMF33, then the mean of the correlation coefficients of the key intrinsic mode function sequence numbers of any two degradation performance factors is the mean of the correlation coefficients of the key intrinsic mode function sequence number IMF11 of the degradation performance factor B1 and the key intrinsic mode function sequence number IMF21 of the degradation performance factor B2, the correlation coefficients of the key intrinsic mode function sequence number IMF11 of the degradation performance factor B1 and the key intrinsic mode function sequence number IMF33 of the degradation performance factor B3, and the correlation coefficients of the key intrinsic mode function sequence number IMF21 of the degradation performance factor B2 and the key intrinsic mode function sequence number IMF33 of the degradation performance factor B3, wherein the key intrinsic mode function sequence number IMF11 of the degradation performance factor B1 indicates that, after the value sequence corresponding to the degradation performance factor B1 is subjected to variational modal decomposition, different intrinsic mode functions (Intrinsic Mode Function, IMF) are extracted, arranged in descending order of frequency, and the first intrinsic mode function is taken as the key intrinsic mode function of the degradation performance factor B1, the key intrinsic mode function sequence number IMF21 of the degradation performance factor B2 indicates that, after the value sequence corresponding to the degradation performance factor B2 is subjected to variational modal decomposition, different intrinsic mode functions are extracted, arranged in descending order of frequency, and the first intrinsic mode function is taken as the key intrinsic mode function of the degradation performance factor B2, and the key intrinsic mode function sequence number IMF33 of the degradation performance factor B3 indicates that, after the value sequence corresponding to the degradation performance factor B3 is subjected to variational modal decomposition, different intrinsic mode functions are extracted, arranged in descending order of frequency, and the third intrinsic mode function is taken as the key intrinsic mode function of the degradation performance factor B3.

[0078] Specifically, the fitness function can be:

[0079]

[0080] wherein, is the fitness function, and is a preset weight, and is greater than 0, for example, is 0.3, is 0.7, is the correlation coefficient of the mth key intrinsic mode function sequence number of the nth degradation performance factor and the remaining life, is the total number of degradation performance factors, is the total number of key intrinsic mode function sequence numbers of the nth degradation performance factor, is the mean of the correlation coefficient of the intrinsic mode function sequence numbers of any two degradation performance factors included by the individual.

[0081] It can be understood that the above fitness function can ensure that the intrinsic mode function sequence number of the selected degradation performance factor is most relevant to the remaining life of the OPGW optical cable, directly improving the accuracy of life prediction. Reduce the redundancy of intrinsic mode functions between different degradation performance factors, avoid information duplication or collinearity problem, and enhance the robustness of the model.

[0082] Figure 4 is a structural schematic diagram of the remaining life prediction model according to some embodiments of the present specification, as shown in Figure 4 As preferred, the remaining life prediction model includes:

[0083] I. Input layer, used for inputting the key intrinsic mode function of each degradation performance factor included in the degradation performance factor group, input shape: (batch_size, num_factors, num_IMFs, sequence_length), wherein batch_size is batch size, indicating the number of samples processed simultaneously by the remaining life prediction model in one iteration, num_factors is the number of degradation performance factors included in the degradation performance factor group, num_IMFs: The number of key intrinsic mode functions of each degradation performance factor included in the degradation performance factor group, sequence_length: The time step of each key intrinsic mode function;

[0084] II. Convolutional layer, including a plurality of convolutional blocks, used for extracting local time-frequency features of the key intrinsic mode function of each degradation performance factor, including:

[0085] 1. First convolutional block:

[0086] Convolutional layer 1: 64 filters, kernel size = 3, step = 1, activation function ReLU;

[0087] Batch normalization (BN): accelerate training and improve stability;

[0088] Max pooling layer 1: pooling size = 2, stride = 2, reduce feature dimension;

[0089] 2、Second convolutional block:

[0090] Convolutional layer 2: 128 filters, kernel size = 3, stride = 1, activation function ReLU.

[0091] Batch normalization;

[0092] Max pooling layer 2: pooling size = 2, stride = 2;

[0093] Three, feature fusion layer, for fusing the local time-frequency features of the key eigenmode functions of each degradation performance factor, constructing a global degradation feature vector, including:

[0094] Channel attention mechanism (SE module): dynamically adjusting the weights of the key eigenmode functions of different degradation performance factors, highlighting the key degradation features.

[0095] Feature concatenation: concatenating the feature maps of the key eigenmode functions of all degradation performance factors in the channel dimension to form a fusion feature vector;

[0096] Four, fully connected layer, for mapping the global degradation feature vector to the OPGW cable remaining life through nonlinear transformation and stepwise dimension reduction, including:

[0097] Fully connected layer 1: 256 neurons, activation function = ReLU, Dropout = 0.5 (to prevent overfitting).

[0098] Fully connected layer 2: 64 neurons, activation function linear activation function.

[0099] Output layer: 1 neuron, activation function = linear (directly regress RUL value);

[0100] Five, output layer, for outputting the OPGW cable remaining life.

[0101] The loss function for training the remaining life prediction model is mean squared error or mean absolute error. Optimizer: choose Adam or SGD with Momentum, initial learning rate set to 1e-3~1e-4. Training includes forward propagation phase, loss calculation and back propagation. Evaluate model performance on validation set (calculate validation loss), monitor whether overfitting. Visualize the scatter plot of predicted OPGW cable remaining life and true OPGW cable remaining life, check systematic bias. If the validation loss does not decrease for N consecutive rounds, terminate training early.

[0102] Input the key proper orthogonal decomposition function of each degradation performance factor included in the extracted degradation performance factor group into the residual life prediction model corresponding to the degradation performance factor group, and output the predicted residual life of the reference OPGW optical cable.

[0103] Figure 5 is an effect diagram of the residual life prediction model according to some embodiments of the present specification, wherein the structure of the comparative model is similar to that of the residual life prediction model, and the input of the comparative model includes each proper orthogonal decomposition function of each degradation performance factor included in the degradation performance factor group. As shown in Figure 5 , compared with the comparative model, the output result of the residual life prediction model is closer to the true value, because the input of the comparative model is not dimensionally reduced, resulting in a large number of input features, which may contain redundant or collinear information, while the input of the residual life prediction model directly focuses on the features strongly related to the life, avoiding the interference of irrelevant or weakly related proper orthogonal decomposition functions, thereby improving the prediction accuracy.

[0104] Step 6, based on the residual life of the reference OPGW optical cable corresponding to each degradation performance factor group, determine the residual life of the OPGW optical cable.

[0105] Specifically includes:

[0106] Based on the performance test data of a plurality of sample OPGW optical cables, a polynomial multiple regression model is established, wherein the independent variable of the polynomial multiple regression model is the residual life of the reference OPGW optical cable corresponding to each degradation performance factor group, and the dependent variable of the polynomial multiple regression model is the residual life of the OPGW optical cable. Specifically, the least squares method or the regularization method (such as Ridge / Lasso regression) can be used to fit the coefficients, and the optimal model order is selected through cross-validation;

[0107] Through the polynomial multiple regression model, based on the residual life of the reference OPGW optical cable corresponding to each degradation performance factor group, determine the residual life of the OPGW optical cable.

[0108] Figure 6 is an effect diagram of the polynomial multiple regression model according to some embodiments of the present specification, as shown in Figure 6 , compared with directly taking the mean of the residual life of the reference OPGW optical cable corresponding to each degradation performance factor group, the output result of the polynomial multiple regression model is closer to the true value, because the polynomial multiple regression model overcomes the linear assumption limitation and information utilization problem of directly taking the mean through nonlinear modeling and multiple factor synergistic optimization, thereby more accurately capturing the complex relationship between the residual life of the reference OPGW optical cable corresponding to each degradation performance factor group and the true residual life.

[0109] Figure 7is a module schematic diagram of an OPGW optical cable life prediction system according to some embodiments of the present specification, as shown in Figure 7 An OPGW optical cable life prediction system can include a factor determination module, a factor grouping module, a data acquisition module, and a life prediction module, as shown in

[0110] The factor determination module is configured to determine a multi-dimensional performance degradation performance factor set, wherein the multi-dimensional performance degradation performance factor set includes a plurality of mechanical degradation performance factors, electrical degradation performance factors, and transmission degradation performance factors.

[0111] The factor grouping module is configured to determine a plurality of degradation performance factor groups based on the multi-dimensional performance degradation performance factor set.

[0112] The data acquisition module is configured to acquire multi-dimensional performance data of the OPGW optical cable based on the multi-dimensional performance degradation performance factor set.

[0113] The life prediction module is configured to, for each degradation performance factor group, establish a residual life prediction model corresponding to the degradation performance factor group, for each degradation performance factor group, extract a performance data set corresponding to the degradation performance factor group from the multi-dimensional performance data of the OPGW optical cable, decompose the performance data set corresponding to the degradation performance factor group by a variational mode decomposition algorithm to obtain a decomposition result, predict a residual life of a reference OPGW optical cable based on the decomposition result by the residual life prediction model corresponding to the degradation performance factor group, and determine the residual life of the OPGW optical cable based on the residual life of the reference OPGW optical cable corresponding to each degradation performance factor group.

[0114] An OPGW optical cable life prediction system can be used to perform an OPGW optical cable life prediction method. For more description of an OPGW optical cable life prediction system, please refer to the description of an OPGW optical cable life prediction method, which will not be repeated here.

[0115] Finally, it should be understood that the embodiments described in the specification are only used to illustrate the principles of the embodiments of the specification. Other variations can also be within the scope of the specification. Therefore, as an example but not limitation, alternative configurations of the embodiments of the specification can be considered consistent with the teachings of the specification. Accordingly, the embodiments of the specification are not limited to the embodiments explicitly introduced and described in the specification.

Claims

1. A method for predicting the lifetime of OPGW optical cables, characterized in that, include: A multidimensional performance degradation performance factor set is determined, wherein the multidimensional performance degradation performance factor set includes multiple mechanical degradation performance factors, electrical degradation performance factors, and transmission degradation performance factors; Based on a multidimensional set of performance degradation factors, multiple groups of degradation factors are identified. For each degradation performance factor group, establish a remaining life prediction model corresponding to the degradation performance factor group; Based on a multidimensional performance degradation factor set, multidimensional performance data of OPGW optical cables are collected. For each degradation performance factor group, the performance dataset corresponding to the degradation performance factor group is extracted from the multidimensional performance data of the OPGW optical cable. The performance dataset corresponding to the degradation performance factor group is decomposed by the variational mode decomposition algorithm to obtain the decomposition result. Based on the decomposition result, the remaining lifetime of the reference OPGW optical cable is predicted by the remaining lifetime prediction model corresponding to the degradation performance factor group. The remaining lifetime of the OPGW optical cable is determined based on the remaining lifetime of the reference OPGW optical cable corresponding to each degradation performance factor group, specifically including: Based on the performance test data of multiple sample OPGW optical cables, a multinomial multiple regression model was established. The independent variable of the multinomial multiple regression model is the remaining life of the reference OPGW optical cable corresponding to each degradation performance factor group, and the dependent variable of the multinomial multiple regression model is the remaining life of the OPGW optical cable. The least squares method or regularization method was used to fit the system, and the optimal model order was selected through cross-validation. The remaining lifespan of the OPGW optical cable is determined by using a multinomial regression model based on the remaining lifespan of the reference OPGW optical cable corresponding to each degradation performance factor group.

2. The method for predicting the lifetime of an OPGW optical cable according to claim 1, characterized in that, Determine the set of multidimensional performance degradation factors, including: Identify multiple degenerative performance factors to be screened; Based on multiple degradation performance factors to be screened, performance test data of multiple sample OPGW optical cables were obtained, among which the remaining lifetime of multiple sample OPGW optical cables differed; Based on the performance test data of multiple sample OPGW optical cables, the correlation coefficient between each degradation performance factor to be screened and the remaining lifetime was determined; Based on the correlation coefficient between each degradation performance factor to be screened and the remaining lifespan, multiple effective degradation performance factors are screened from multiple degradation performance factors to be screened. Based on the performance test data of multiple sample OPGW optical cables, determine the correlation coefficient of any two valid degradation performance factors; For each valid degradation performance factor, the redundancy value of the valid degradation performance factor is calculated based on the correlation coefficient between the valid degradation performance factor and any other valid degradation performance factor. Based on the correlation coefficient of any two valid degenerative performance factors, multiple valid degenerative performance factors are clustered to generate multiple factor clusters; Based on the redundancy value of each valid degradation performance factor and multiple factor clusters, a multidimensional performance degradation performance factor set is determined, wherein the multidimensional performance degradation performance factor set includes multiple degradation performance factors.

3. The method for predicting the lifetime of an OPGW optical cable according to claim 2, characterized in that, Based on the redundancy value of each valid degradation performance factor and multiple factor clusters, a multidimensional performance degradation performance factor set is determined, including: Establish factor sampling constraints and cluster coverage constraints; Establish factor sampling constraints and cluster coverage constraints, sample multiple factor clusters, and generate the first generation population; Establish a fitness function, which is related to the redundancy of the effective degenerative performance factors included in the individual and the correlation coefficient between any two effective degenerative performance factors; For each individual, determine the individual's coverage of multiple factor clusters, and calculate the individual's fitness value based on the fitness function; Based on each individual's coverage and fitness values ​​across multiple factor clusters, several individuals are selected for retention. For each retained individual, determine the factor cluster coverage vector of the retained individual; Calculate the cosine similarity of the factor cluster coverage vectors of any two retained individuals; Cross-reference multiple retained individuals based on the cosine similarity of the factor cluster coverage vectors of any two retained individuals; Based on the coverage of multiple factor clusters by each retained individual, multiple retained individuals are mutated; Based on the crossover and mutation results, update the population and perform iterative optimization until the first termination condition is met.

4. The method for predicting the lifetime of an OPGW optical cable according to claim 2, characterized in that, Based on a multidimensional set of performance degradation factors, several groups of degradation factors were identified, including: Using a genetic algorithm, multiple groups of degenerative performance factors are determined based on the correlation coefficient between any two degenerative performance factors.

5. A method for predicting the lifetime of an OPGW optical cable according to any one of claims 1-4, characterized in that, Based on a multidimensional performance degradation factor set, multidimensional performance data of OPGW optical cables are collected, including at least: The attenuation coefficient of OPGW optical cable is obtained using optical time-domain reflectometry. The polarization state of OPGW optical cable is obtained using polarized light time-domain reflectometry. The resistance of the metal conductors in OPGW optical cables was measured using the four-terminal method. The dispersion parameters of OPGW optical cables were measured using an optical frequency domain reflectometer. The grounding resistance value of the OPGW optical cable was measured using a grounding resistance tester.

6. A method for predicting the lifetime of an OPGW optical cable according to any one of claims 2-4, characterized in that, The performance dataset corresponding to the degradation performance factor group is decomposed using the variational mode decomposition algorithm, and the decomposition results are obtained, including: Based on the performance test data of multiple sample OPGW optical cables and the variational mode decomposition algorithm, the key intrinsic mode function index of each degradation performance factor included in the degradation performance factor group is determined; The performance dataset corresponding to the degradation performance factor group is decomposed by the variational mode decomposition algorithm to obtain multiple intrinsic mode functions of each degradation performance factor included in the degradation performance factor group. Based on the key intrinsic mode function index of each degenerate performance factor included in the degenerate performance factor group, the key intrinsic mode function of each degenerate performance factor included in the degenerate performance factor group is extracted from the multiple intrinsic mode functions of each degenerate performance factor included in the degenerate performance factor group.

7. The method for predicting the lifetime of an OPGW optical cable according to claim 6, characterized in that, Based on performance test data from multiple sample OPGW optical cables and variational mode decomposition algorithms, the key intrinsic mode function indices for each degradation performance factor in the degradation performance factor group were determined, including: Based on the performance test data of multiple sample OPGW optical cables and the variational mode decomposition algorithm, the correlation coefficient between the intrinsic mode function index of each degradation performance factor and the remaining lifetime, as well as the correlation coefficient between the intrinsic mode function indexes of any two degradation performance factors, are determined. Based on the correlation coefficient between the intrinsic mode function index of each degradation performance factor and the remaining lifetime, and the correlation coefficient between the intrinsic mode function indexes of any two degradation performance factors, the key intrinsic mode function index of each degradation performance factor included in the degradation performance factor group is determined.

8. The method for predicting the lifetime of an OPGW optical cable according to claim 7, characterized in that, The remaining lifetime prediction model includes: The input layer is used to input the key intrinsic mode functions of each degradation performance factor included in the degradation performance factor group; Convolutional layers, comprising multiple convolutional blocks, are used to extract local time-frequency features of the key intrinsic mode functions for each degraded performance factor; The feature fusion layer is used to fuse the local time-frequency features of the key intrinsic mode functions of each degradation performance factor to construct a global degradation feature vector; A fully connected layer is used to map the global degradation feature vector to the remaining lifetime of the OPGW optical cable through nonlinear transformation and stepwise dimensionality reduction; The output layer is used to output the remaining lifespan of the OPGW optical cable.

9. An OPGW optical cable lifetime prediction system, characterized in that, A method for predicting the lifetime of an OPGW optical cable according to any one of claims 1-8 includes: A factor determination module is used to determine a multidimensional performance degradation performance factor set, wherein the multidimensional performance degradation performance factor set includes multiple mechanical degradation performance factors, electrical degradation performance factors, and transmission degradation performance factors; The factor grouping module is used to determine multiple degradation performance factor groups based on a multidimensional set of performance degradation performance factors. The data acquisition module is used to collect multidimensional performance data of OPGW optical cables based on a multidimensional performance degradation factor set. The lifetime prediction module is used to establish a remaining lifetime prediction model for each degradation performance factor group. For each degradation performance factor group, the performance dataset corresponding to the degradation performance factor group is extracted from the multidimensional performance data of the OPGW optical cable. The performance dataset corresponding to the degradation performance factor group is decomposed by the variational mode decomposition algorithm to obtain the decomposition result. Based on the decomposition result, the remaining lifetime of the reference OPGW optical cable is predicted by the remaining lifetime prediction model corresponding to the degradation performance factor group. Based on the remaining lifetime of the reference OPGW optical cable corresponding to each degradation performance factor group, the remaining lifetime of the OPGW optical cable is determined.

Images