A method and device for predicting the life of an optical fiber, a terminal and a storage medium thereof

Abstract

The application particularly relates to a kind of optical fiber life prediction method, device, terminal and its storage medium, the prediction method includes: with service life, factory parameter and environmental parameter as data input, form the original data set of to-be-measured optical fiber;The original data set is input into optical fiber fatigue strength prediction model, obtains prediction data set;The prediction data set at least includes fatigue parameter and wilb distribution slope;The prediction data set is input into optical fiber life prediction model, obtains the residual life value of to-be-measured optical fiber.The optical fiber fatigue strength prediction model and the optical fiber life prediction model are connected in series in the application, the output of optical fiber fatigue strength prediction model is used as the input of optical fiber life prediction model, a variety of factors influencing the life of optical fiber are considered comprehensively, and the accuracy of residual life prediction of to-be-measured optical fiber is improved by combining the method of machine learning.

Description

Technical Field

[0001] This invention relates to the fields of power and communication technology, and specifically to a method, apparatus, terminal and storage medium for predicting the lifetime of optical fibers. Background Technology

[0002] Optical fiber communication technology is one of the main pillars of modern communication, a significant symbol of the new technological revolution, and a primary means of transmitting various types of information in the future information society. The reliability of the optical fiber cable, its transmission carrier, directly affects the communication quality of optical fiber communication. my country has been developing optical fiber cables for over 40 years, and its production technology is highly mature. The amount of optical fiber cable laid has been increasing year by year, and the aging problem of optical fibers has gradually become a priority. Among these, the assessment of the reliability and lifespan of optical fibers has gradually become a key research focus for domestic project teams.

[0003] Existing technologies for predicting fiber optic lifetime mostly focus on single indicators such as temperature and stress, or the parameters set in the experiment are too ideal, resulting in large prediction errors and making it difficult to guarantee the reliability of fiber optic lifetime prediction. Summary of the Invention

[0004] This invention addresses the technical problems in existing optical fiber lifetime prediction technologies, such as simplistic modeling and overly ideal parameter settings leading to excessively large prediction errors. It provides an optical fiber lifetime prediction method, device, terminal, and storage medium, which at least have the advantages of comprehensive consideration and high accuracy.

[0005] First aspect

[0006] This invention provides a method for predicting the lifetime of an optical fiber, comprising:

[0007] Using the service life, environmental parameters, and test data as inputs, a raw dataset of the optical fiber under test is formed; the raw dataset is then input into the optical fiber fatigue strength prediction model to obtain performance parameters; the performance parameters include at least fatigue parameters and Welbuck distribution slope.

[0008] Using the factory parameters and the aforementioned performance parameters as data inputs, a prediction dataset for the optical fiber under test is formed. The prediction dataset is then input into the optical fiber lifetime prediction model to obtain the remaining lifetime value of the optical fiber under test.

[0009] The fiber fatigue strength prediction model is a model trained based on the performance parameters corresponding to multiple fatigue sample datasets; the service life, environmental parameters and test data included in any fatigue sample dataset correspond to the fatigue parameters and Welb distribution slope included in the performance parameters.

[0010] The fiber optic lifetime prediction model is a model trained based on the remaining lifetime values ​​corresponding to multiple prediction datasets. The factory parameters and performance parameters included in any of the prediction datasets correspond to the remaining lifetime value.

[0011] Specifically, one of the main technical concepts of this invention is to connect the optical fiber fatigue strength prediction model and the optical fiber lifetime prediction model in series, use the output of the optical fiber fatigue strength prediction model as the input of the optical fiber lifetime prediction model, comprehensively consider various factors affecting the optical fiber lifetime, and combine machine learning methods to improve the accuracy of the remaining lifetime prediction of the optical fiber under test.

[0012] Furthermore, the fatigue parameters include the following calculation steps:

[0013] S10. Divide the optical fiber to be tested into multiple optical fiber samples with a length of S and a number of segments of N.

[0014] S11. Pre-set constant strain rate, temperature, and relative humidity during the test;

[0015] S12. Record the breaking tension T of all optical fiber samples from the 1st to the Nth fiber.

[0016] S13. Calculate the fracture stress of all fiber optic samples: σ f =T / A g ;σ f The tensile stress is represented by T, and the tensile force is represented by Ag, which is the nominal cross-sectional area of ​​the glass optical fiber.

[0017] S14. Calculate the stress rate of all the optical fibers under test: Represents the stress rate, t(σ) f ) represents the fracture time, t(0.8σ f This indicates the time taken to reach 80% of the fracture stress;

[0018] S15. Calculate the fatigue parameters of the optical fiber under test: σ f (0.5) represents the fracture stress when the cumulative failure probability is 0.5. nd Indicates fatigue parameters, σ1 This represents the fracture stress at a stress rate of 1.

[0019] Specifically, another concept of this invention is to integrate the theory of brittle material fracture into the machine learning process, and to fit the service life, environmental parameters and actual test results together to form a transition from reality to model calculation. In this way, based on real-world experiments, the reliability of model calculation is improved, thereby improving the fitting accuracy of the remaining life value.

[0020] Furthermore, the slope of the Welb distribution includes the following calculation steps:

[0021] S20. Rearrange all the optical fibers to be tested in order of their fracture stress, from lowest to highest as 1, 2, ..., N;

[0022] S21. Calculate the cumulative breakage probability of each optical fiber under test: F i = (i-0.5) / N; F i This represents the cumulative fracture probability, where i takes values ​​of 1, 2, ..., N;

[0023] S22. Plot ln[-ln(1-F)] i )] for ln(σ f The Weibull curve is plotted, and data is marked on the curve.

[0024] S23. Based on the dynamic Weibull curve and cumulative function Calculate the Welb slope: m d =2.46 / ln[σ f (0.85)]-ln[σ f (0.15)];m d Let σ0 represent the slope of the Weibull distribution, and σ0 represent the Weibull parameter. f (0.85) represents the fracture stress when the cumulative failure probability is 0.85, σ f (0.15) represents the fracture stress when the cumulative failure probability is 0.15.

[0025] Furthermore, the remaining lifetime value includes the following calculation steps:

[0026] S30. Based on the predicted dataset, calculate the remaining lifetime of the optical fiber under test:

[0027] t f This indicates the remaining lifetime of the optical fiber under test. σ represents the adjustment factor. p Indicates screening stress, t p Indicates the effective screening time, F represents the fiber failure rate, and N represents the effective screening time. p σ represents the average breakage rate per unit length obtained from the fiber optic factory screening test. a N represents the static stress on the optical fiber, n represents the actual fatigue parameter of the optical fiber, and N represents the static stress on the optical fiber. p This represents the microcrack propagation rate, and L represents the equivalent tensile length. It's worth noting that n represents the actual fatigue parameter of the optical fiber, i.e., the actual n... d value.

[0028] Optionally, the service life is t. i ;

[0029] The factory parameters include σ p t p F, N p σ a N p The specific values ​​of L and L are provided by the supplier of the optical fiber to be tested.

[0030] The environmental parameters include atmospheric pressure, maximum operating temperature, minimum operating temperature, operating temperature difference, maximum operating wind speed level, minimum operating wind speed level, average operating wind speed level, and number of lightning strikes.

[0031] The number of lightning strikes is an independent value, denoted as Y; the characteristic values ​​of the environmental parameters other than the number of lightning strikes are denoted as X. j j = 1, 2, ..., 7.

[0032] Specifically, the formulas used in the model include at least actual data affecting the remaining lifespan of the optical fiber, such as service life, factory parameters, and environmental parameters. Multiple real-world factors affecting the lifespan of the optical fiber under test are coupled into the optical fiber fatigue strength prediction model and the optical fiber lifespan prediction model, thereby increasing the weight of the influence of real-world factors on the remaining lifespan value of the optical fiber under test and improving the prediction accuracy of the present invention.

[0033] Optionally, the performance parameters include median fracture time and maximum strain of the optical fiber;

[0034] The formula for calculating the median fracture time is as follows:

[0035] The formula for calculating the maximum strain of the optical fiber is: σ=E0(1+cε)ε, where σ represents the stress on the optical fiber, E0 represents Young's modulus, c represents the material correlation coefficient with a value of 2.5, and ε represents the maximum strain of the optical fiber.

[0036] Optionally, the fiber fatigue strength prediction model and / or the fiber lifetime prediction model are constructed using the following steps:

[0037] S40. Feature extraction is performed using a convolutional neural network;

[0038] S41. Learn the feature relationships on the time series through a long short-term memory network;

[0039] S42. Form a model network.

[0040] Second aspect

[0041] This invention provides an optical fiber lifetime prediction apparatus for implementing an optical fiber lifetime prediction method as described in any embodiment of the first aspect, comprising:

[0042] The generation unit is used to generate the original dataset based on the service life, environmental parameters, and test data of the optical fiber under test; and to generate the predicted dataset based on the factory parameters and performance parameters of the optical fiber under test.

[0043] The processing unit is used to input the original dataset into the optical fiber fatigue strength prediction model to obtain performance parameters; and to input the predicted dataset into the optical fiber lifetime prediction model to output the remaining lifetime value of the optical fiber under test.

[0044] Specifically, another key concept of this invention lies in constructing an optical fiber lifetime prediction device based on the optical fiber lifetime prediction method provided in any embodiment of the first aspect. The generation unit processes data, converting actual and calculated data such as service life, factory parameters, environmental parameters, test data, and performance parameters into a dataset that can be processed by the optical fiber fatigue strength prediction model and the optical fiber lifetime prediction model. The processing unit performs machine learning based on the dataset, supported by the optical fiber fatigue strength prediction model, to obtain the remaining lifetime value of the optical fiber under test. The remaining lifetime value of the optical fiber under test can be predicted through the aforementioned generation and processing units.

[0045] It is worth explaining that the optical fiber fatigue strength prediction model is a model trained on the prediction datasets corresponding to multiple fatigue sample datasets respectively; the service life, environmental parameters and test data included in any of the fatigue sample datasets correspond to the fatigue parameters and Welb distribution slope included in the performance parameters; the optical fiber lifetime prediction model is a model trained on the remaining lifetime values ​​corresponding to multiple prediction datasets respectively, and the factory parameters and performance parameters included in any of the prediction datasets correspond to the remaining lifetime value.

[0046] Third aspect

[0047] The present invention also provides a terminal comprising at least one processor, a communication interface, and a memory, wherein the communication interface is used to send and / or receive data, the memory is used to store a computer program, and the at least one processor is used to invoke the computer program stored in the at least one memory to implement a fiber optic lifetime prediction method as provided in any embodiment of the first aspect.

[0048] It should be noted that the processor included in the terminal described in the fifth aspect above can be a processor specifically designed to execute these methods (referred to as a dedicated processor for easy distinction), or a processor that executes these methods by calling a computer program, such as a general-purpose processor. Optionally, at least one processor may include both dedicated and general-purpose processors.

[0049] Optionally, the computer program described above can be stored in memory. For example, the memory can be a non-transitory memory, such as read-only memory (ROM), which can be integrated with the processor on the same device or disposed on different devices. The embodiments of the present invention do not limit the type of memory or the arrangement of the memory and the processor.

[0050] In one possible implementation, at least one of the aforementioned memories is located outside the aforementioned terminal.

[0051] In yet another possible implementation, at least one of the aforementioned memories is located within the aforementioned terminal.

[0052] In another possible implementation, a portion of the memory of the at least one memory is located inside the terminal, while another portion of the memory is located outside the terminal.

[0053] In this invention, the processor and memory may also be integrated into a single device, that is, the processor and memory can be integrated together.

[0054] Fourthly, embodiments of the present invention provide a server, which includes a processor, a memory, and a communication interface; the memory stores a computer program; when the processor executes the computer program, the communication interface is used to send and / or receive data, and the server can perform a fiber optic lifetime prediction method as provided in any embodiment of the first aspect.

[0055] It should be noted that the processor included in the server described in the fourth aspect above can be a processor specifically designed to execute these methods (referred to as a dedicated processor for distinction), or a processor that executes these methods by calling computer programs, such as a general-purpose processor. Optionally, at least one processor may include both dedicated and general-purpose processors.

[0056] Optionally, the computer program described above can be stored in memory. For example, the memory can be a non-transitory memory, such as read-only memory (ROM), which can be integrated with the processor on the same device or disposed on different devices. The embodiments of the present invention do not limit the type of memory or the arrangement of the memory and the processor.

[0057] In one possible implementation, at least one of the aforementioned storage devices is located outside the aforementioned server.

[0058] In yet another possible implementation, at least one of the aforementioned storage devices is located within the aforementioned server.

[0059] In another possible implementation, a portion of the memory of the at least one memory is located within the server, while another portion of the memory is located outside the server.

[0060] In this invention, the processor and memory may also be integrated into a single device, that is, the processor and memory can be integrated together.

[0061] Fifth aspect

[0062] The present invention also provides a computer-readable storage medium storing a computer program that, when run on a processor, implements a method for predicting the lifetime of an optical fiber as described in any embodiment of the first aspect.

[0063] In a sixth aspect, the present invention provides a computer program product comprising a computer program that, when run on at least one processor, implements a method for predicting the lifetime of an optical fiber as described in any embodiment of the first aspect.

[0064] Optionally, the computer program product can be a software installation package, which can be downloaded and executed on a computing device when the aforementioned method is required.

[0065] The beneficial effects of the technical solutions provided in the second to sixth aspects of this invention can be referred to the beneficial effects of the technical solutions provided in the first aspect, and will not be repeated here.

[0066] In summary, the present invention provides a method, apparatus, terminal, and storage medium for predicting optical fiber lifetime, which has at least the following advantages:

[0067] 1. This invention connects the optical fiber fatigue strength prediction model and the optical fiber lifetime prediction model in series, uses the output of the optical fiber fatigue strength prediction model as the input of the optical fiber lifetime prediction model, comprehensively considers various factors affecting optical fiber lifetime, and combines machine learning methods to improve the accuracy of predicting the remaining lifetime of the optical fiber under test.

[0068] 2. This invention integrates the theory of brittle material fracture into the machine learning process, and fits the service life, environmental parameters and test data obtained from real-world experiments to form a transition from reality to model calculation. Based on the existing service life and environmental parameters, and relying on real-world experiments, the reliability of model calculation is improved, thereby improving the fitting accuracy of the remaining life value.

[0069] 3. The formulas used in the optical fiber fatigue strength prediction model and optical fiber life prediction model provided by the present invention include at least actual data affecting the remaining life of the optical fiber, such as service life, factory parameters and environmental parameters. The various real factors affecting the life of the optical fiber under test are coupled into the optical fiber fatigue strength prediction model and optical fiber life prediction model, thereby strengthening the influence weight of the actual factors on the remaining life value of the optical fiber under test and improving the prediction accuracy of the present invention. Attached Figure Description

[0070] The present invention will be further described in detail below with reference to the accompanying drawings and preferred embodiments. However, those skilled in the art will understand that these drawings are drawn only for the purpose of explaining the preferred embodiments and therefore should not be construed as limiting the scope of the invention. Furthermore, unless specifically indicated, the drawings are only schematic representations of the composition or structure of the described objects and may contain exaggerated depictions, and the drawings are not necessarily drawn to scale.

[0071] Figure 1 A schematic diagram of a method for predicting the lifetime of an optical fiber provided in an embodiment of the present invention;

[0072] Figure 2 A schematic diagram of the processing steps of a method for predicting the lifetime of an optical fiber provided in another embodiment of the present invention;

[0073] Figure 3 A schematic diagram of the structure of a terminal provided in one embodiment of the present invention;

[0074] Figure 4 A schematic diagram of the structure of an optical fiber lifetime prediction device provided in one embodiment of the present invention;

[0075] Figure 5 A schematic diagram of the structure of a fiber lifetime prediction training device provided in one embodiment of the present invention;

[0076] Figure 6 This invention provides a schematic diagram of the interaction between a fiber optic lifetime terminal and a server. Detailed Implementation

[0077] The following is in conjunction with the appendix Figures 1 to 6 The present invention will be described in detail below.

[0078] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0079] The main concept of this invention lies in comprehensively considering various factors affecting the lifespan of optical fibers, such as their service life, manufacturing parameters, environmental parameters, and operational performance parameters. It generates fiber fatigue strength prediction models and fiber lifespan prediction models through deep learning, thereby predicting the remaining lifespan of the tested optical fiber based on actual service life, manufacturing parameters, environmental parameters, and performance parameters. This solves the technical problem of insufficient consideration of influencing factors in fiber lifespan prediction in existing technologies. Furthermore, by combining the power law of brittle material fracture theory, sample data is directly obtained through mechanical stress experiments, making the prediction of the remaining lifespan of the tested optical fiber more accurate. Therefore, this invention can effectively predict the remaining lifespan of optical fibers, with higher prediction accuracy and faster efficiency compared to existing technologies. This helps to detect potential fiber optic failure risks in advance and reduces the probability of fiber optic network outages.

[0080] For an explanation and description of the technical solution provided by this invention, please refer to [link / reference]. Figure 1 The diagram shown is a schematic representation of a fiber lifetime prediction method provided in an embodiment of the present invention.

[0081] Specifically, the main concept of this invention lies in providing a fiber optic lifetime prediction method that combines a fiber fatigue strength prediction model and a fiber lifetime prediction model. The fiber fatigue strength prediction model comprehensively considers the factors influencing fiber lifetime, encompassing multiple factors affecting fiber lifetime in the form of an original dataset. It utilizes machine learning to calculate and fit the data in the original dataset, outputting a predicted dataset for calculating the remaining lifetime value of the fiber. The fiber lifetime prediction model, on the other hand, calculates and fits the predicted dataset, outputting the remaining lifetime value of the fiber under test. Since the predicted dataset is generated based on the original dataset, which includes the service life of the fiber under test, environmental parameters, and test data—actual parameters closely related to the fiber lifetime—the data in the predicted dataset calculated from the original dataset is more accurate. Furthermore, the predicted dataset includes factory parameters and performance parameters. The fiber lifetime prediction model calculates and fits the remaining lifetime value of the fiber under test using brittle material fracture theory. Thus, supported by the fiber fatigue strength prediction model, multiple factors affecting fiber lifetime are comprehensively considered, resulting in a more accurate remaining lifetime value obtained through the fiber lifetime prediction model.

[0082] It is worth explaining that both the fiber fatigue strength prediction model and the fiber lifetime prediction model are obtained through machine learning. The fiber fatigue strength prediction model is a model trained on the prediction datasets corresponding to multiple fatigue sample datasets; the service life, environmental parameters, and test data included in any of the fatigue sample datasets correspond to the fatigue parameters and Welb distribution slopes included in the performance parameters; the fiber lifetime prediction model is a model trained on the remaining lifetime values ​​corresponding to multiple prediction datasets; the factory parameters and performance parameters included in any of the prediction datasets correspond to the remaining lifetime values.

[0083] It is worth noting that the technical concept of connecting the optical fiber fatigue strength prediction model and the optical fiber lifetime prediction model in series in this invention is that the two have different parameter outputs. In order to improve the accuracy of lifetime prediction, this invention isolates the two models from each other to reduce crosstalk during optical fiber lifetime prediction, thereby improving prediction accuracy.

[0084] Furthermore, Figure 2 A schematic diagram of the processing steps of a method for predicting the lifetime of an optical fiber provided in another embodiment of the present invention.

[0085] Specifically, since this invention employs an optical fiber fatigue strength prediction model and an optical fiber lifetime prediction model to predict the remaining lifetime of the optical fiber, the prediction method of this invention includes three steps set in sequence:

[0086] 1. Using the service life, factory parameters, and environmental parameters as data inputs, the raw dataset of the optical fiber under test is formed;

[0087] 2. Input the original dataset into the optical fiber fatigue strength prediction model to obtain the prediction dataset;

[0088] 3. Input the predicted dataset into the optical fiber lifetime prediction model to obtain the remaining lifetime value of the optical fiber under test.

[0089] Since the remaining lifetime value of the optical fiber under test can be obtained through only the above three steps, the efficiency of predicting the remaining lifetime value of the optical fiber under test can be greatly improved. Moreover, both the optical fiber fatigue strength prediction model and the optical fiber remaining lifetime prediction model are obtained through continuous training by machine learning, which ensures the accuracy of the prediction of the remaining lifetime value of the optical fiber under test.

[0090] Furthermore, the optical fiber fatigue strength prediction model is created using microcrack texture theory and provides a method for calculating fatigue parameters and Welb distribution slope.

[0091] The fatigue parameters include the following calculation steps:

[0092] S10. Divide the optical fiber to be tested into multiple optical fiber samples with a length of S and a number of segments of N.

[0093] S11. Pre-set constant strain rate, temperature, and relative humidity during the test;

[0094] S12. Record the breaking tension T of all optical fiber samples from the 1st to the Nth fiber.

[0095] S13. Calculate the fracture stress of all fiber optic samples: σ f =T / A g ;σ f The tensile stress is represented by T, and the tensile force is represented by Ag, which is the nominal cross-sectional area of ​​the glass optical fiber.

[0096] S14. Calculate the stress rate of all the optical fibers under test: Represents the stress rate, t(σ) f ) represents the fracture time, t(0.8σ f This indicates the time taken to reach 80% of the fracture stress;

[0097] S15. Calculate the fatigue parameters of the optical fiber under test: σ f (0.5) represents the fracture stress when the cumulative failure probability is 0.5. nd Indicates fatigue parameters, σ1 This represents the fracture stress at a stress rate of 1.

[0098] Specifically, the fatigue parameter calculation data originates from actual experiments, i.e., test data, which can effectively improve the accuracy of the fiber optic fatigue strength prediction model. It's worth understanding that the fatigue sample dataset is obtained through real-world experiments; that is, the training of the fiber optic fatigue strength prediction model is supported by real-world experimental data, making the trained fiber optic fatigue strength prediction model grounded in reality. For example:

[0099] The test optical fibers are 5m-15m long and consist of 30-100 fibers; preferably S=10m and N=50.

[0100] The constant strain rate of the test optical fiber is 10%-30% / min, the temperature is 0℃-40℃, and the relative humidity is 30%-70%, preferably a constant strain rate of 15%, a temperature of 20℃, and a relative humidity of 50%.

[0101] After completing the parameter settings, start the testing machine to stretch the optical fiber, record the stress-time relationship curve until the optical fiber breaks, and then stop the motor. This will allow you to obtain the breaking tension T of the first to Nth optical fiber samples.

[0102] Then, the fatigue parameters of the optical fiber are calculated according to steps S13 to S15. nd .

[0103] The slope of the Welb distribution is calculated using the following steps:

[0104] S20. Rearrange all the optical fibers to be tested in order of their fracture stress, from lowest to highest as 1, 2, ..., N;

[0105] S21. Calculate the cumulative breakage probability of each optical fiber under test: F i = (i-0.5) / N; F i This represents the cumulative fracture probability, where i takes values ​​of 1, 2, ..., N;

[0106] S22. Plot ln[-ln(1-F)] i )] for ln(σ f The Weibull curve is plotted, and data is marked on the curve.

[0107] S23. Based on the dynamic Weibull curve and cumulative function Calculate the Welb slope: m d =2.46 / ln[σ f (0.85)]-ln[σ f (0.15)];m d Let σ0 represent the slope of the Weibull distribution, σ0 represent the Weibull parameter, and σf(0.85) represent the fracture stress when the cumulative failure probability is 0.85. f (0.15) represents the fracture stress when the cumulative failure probability is 0.15.

[0108] Specifically, the slope of the Welb distribution is calculated based on fatigue parameters.

[0109] Furthermore, the remaining lifetime value is confirmed using the power law of brittle material fracture theory, including the following calculation steps:

[0110] S30. Based on the predicted dataset, calculate the remaining lifetime of the optical fiber under test:

[0111] t f This indicates the remaining lifetime of the optical fiber under test. σ represents the adjustment factor. p Indicates screening stress, t p Indicates the effective screening time, F represents the fiber failure rate, and N represents the effective screening time. p σ represents the average breakage rate per unit length obtained from the fiber optic factory screening test. a N represents the static stress on the optical fiber, n represents the actual fatigue parameter of the optical fiber, and N represents the static stress on the optical fiber. p The value represents the microcrack propagation rate, and L represents the equivalent tensile length.

[0112] It is worth noting that the service life mentioned is t. i ;

[0113] The factory parameters include σ p t p F, N p σ a N p The specific values ​​of L and L are provided by the supplier of the optical fiber to be tested; among which the adjustment factor The value range is 0.9-1.5.

[0114] The environmental parameters include atmospheric pressure, maximum operating temperature, minimum operating temperature, operating temperature difference, maximum operating wind speed level, minimum operating wind speed level, average operating wind speed level, and number of lightning strikes.

[0115] The test data is obtained through sampling and experimentation, and is used to test the actual material parameters of the optical fiber.

[0116] The number of lightning strikes is an independent value, denoted as Y; the characteristic values ​​of the environmental parameters other than the number of lightning strikes are denoted as X. j j = 1, 2, ..., 7.

[0117] The lifespan of optical fibers ranges from 1 to 30 years (the conventional lifespan of optical fibers is 30 years). If an optical fiber has been in operation for 1.5 years, its lifespan is 1.5 years. If an optical fiber has been in operation for 10 years, its lifespan is 10 years.

[0118] The environmental parameters of optical fiber include atmospheric pressure, operating temperature difference (spring, summer, autumn, and winter), operating maximum temperature (spring, summer, autumn, and winter), operating minimum temperature (spring, summer, autumn, and winter), operating average wind speed level (spring, summer, autumn, and winter), operating maximum wind speed level (spring, summer, autumn, and winter), operating minimum wind speed level (spring, summer, autumn, and winter), and lightning strike records.

[0119] The following parameters—atmospheric pressure, ambient temperature difference (spring, summer, autumn, winter), highest ambient temperature (spring, summer, autumn, winter), lowest ambient temperature (spring, summer, autumn, winter), average ambient wind speed (spring, summer, autumn, winter), maximum ambient wind speed (spring, summer, autumn, winter), and minimum ambient wind speed (spring, summer, autumn, winter)—need to be normalized.

[0120] The feature values ​​of lightning strike records need to be one-hot encoded. If no lightning strike has occurred, the value is set to 0; if a lightning strike has occurred, the value is set to 1.

[0121] There are two main normalization methods: zero-mean standardization (Z-score standardization) and min-max normalization. Min-max normalization maps sample values ​​to [0, 1] through a transformation. The standardization method for sample data X(x1, x2, ...) is as follows: i = 1, 2, ...;

[0122] Where x i x represents the original value of a certain feature for each sample. min It is the minimum value of a certain feature in the sample, x. max It is the maximum value of a certain feature in the sample, x i,norm This represents the standardized feature value.

[0123] The performance parameters include median fracture time and maximum strain of the optical fiber;

[0124] The formula for calculating the median fracture time is as follows:

[0125] The formula for calculating the maximum strain of the optical fiber is: σ=E0(1+cε)ε, where σ represents the stress on the optical fiber, E0 represents Young's modulus, c represents the material correlation coefficient, and ε represents the maximum strain of the optical fiber.

[0126] Among them, Young's modulus E0 is 70.3 GPa; c is 2.5.

[0127] Based on the aforementioned calculation formula, multiple fatigue sample datasets are set up with different service life, environmental parameters, and test data. Each sample dataset is tested according to the experimental procedures to obtain a prediction dataset. The remaining lifetime value is then calculated using the prediction dataset, forming a complete training sample that can be used for both fiber fatigue strength prediction and fiber lifetime prediction models. These training samples are then continuously used to refine and improve the fiber fatigue strength prediction model and fiber lifetime prediction model for predicting the remaining lifetime of the fiber under test. In other words, subsequent input only requires the original dataset to obtain the remaining lifetime value of the fiber under test.

[0128] Furthermore, the construction steps of the optical fiber fatigue strength prediction model and / or the optical fiber lifetime prediction model are as follows: S40, feature extraction is performed through a convolutional neural network; S41, the feature relationship on the time series is learned through a long short-term memory network; S42, a model network is formed.

[0129] Optionally, the sample splitting uses the cross-validation method. K-fold cross-validation is the most common form, where K is typically 10. The dataset is divided into K equal subsets (or as equal as possible). Each time, one subset is reserved as the test set, while the remaining K-1 subsets are used as training data. This process is repeated K times, with each subset used as the test set once. The model is evaluated as the average of these K trial results.

[0130] For further details, please see Figure 3 This is a schematic diagram of the structure of a terminal provided in an embodiment of the present invention.

[0131] Specifically, the terminal includes a processor, a communication interface, and a memory. The processor, communication interface, and memory can be connected via a bus or other means; this embodiment of the invention uses a bus connection as an example.

[0132] The processor is the core of the terminal's computing and control system. It can parse various instructions and data within the terminal. For example, the processor can act as a Central Processing Unit (CPU) to transmit various interactive data between internal structures of the terminal. The communication interface can optionally include standard wired interfaces and wireless interfaces (such as Wi-Fi, mobile communication interfaces, etc.), which, under the control of the processor, can be used to send and receive data. The communication interface can also be used for the transmission and interaction of internal signaling or instructions within the terminal. The memory is the storage device in the terminal used to store programs and data. It can be understood that the memory here includes both the terminal's built-in memory and extended memory supported by the terminal. The memory provides storage space, which stores the terminal's operating system and the program code or instructions required by the processor to execute corresponding operations. Optionally, the memory can also store relevant data generated by the processor after executing the corresponding operation.

[0133] In this embodiment of the invention, the processor runs executable program code in memory to perform the following operations:

[0134] Using the service life, environmental parameters, and test data as inputs, a raw dataset of the optical fiber under test is formed; the raw dataset is then input into the optical fiber fatigue strength prediction model to obtain performance parameters; the performance parameters include at least fatigue parameters and Welbuck distribution slope.

[0135] Using the factory parameters and the aforementioned performance parameters as data inputs, a prediction dataset for the optical fiber under test is formed. The prediction dataset is then input into the optical fiber lifetime prediction model to obtain the remaining lifetime value of the optical fiber under test.

[0136] In one alternative, the processor is further configured to:

[0137] The optical fiber fatigue strength prediction model and optical fiber lifetime prediction model are received from the server through the communication interface, and the models stored in the memory are updated.

[0138] In yet another alternative, the processor is further configured to:

[0139] Obtain multiple fatigue sample datasets and multiple prediction datasets corresponding to the multiple fatigue sample datasets, as well as multiple remaining life values ​​corresponding to the multiple prediction datasets.

[0140] A fiber fatigue strength prediction model is obtained by training multiple fatigue sample datasets and multiple performance parameters corresponding to the multiple fatigue sample datasets; and a fiber lifetime prediction model is obtained by training multiple prediction datasets and multiple remaining lifetime values ​​corresponding to the multiple prediction datasets.

[0141] For further details, please see Figure 4 The diagram shown is a structural schematic of an optical fiber lifetime prediction device provided in an embodiment of the present invention.

[0142] Specifically, the fiber lifetime prediction device can be the aforementioned terminal or a device within the terminal. This device includes a generation unit and a processing unit, wherein:

[0143] The generation unit is used to form the original dataset of the optical fiber under test based on the service life, environmental parameters and test data as data input; and to generate the prediction dataset based on the factory parameters and performance parameters of the optical fiber under test.

[0144] The processing unit is used to input the original dataset into the optical fiber fatigue strength prediction model to obtain performance parameters; and to input the predicted dataset into the optical fiber lifetime prediction model to obtain the remaining lifetime value of the optical fiber under test.

[0145] In one possible implementation, the fiber lifetime prediction device further includes:

[0146] The receiving unit is used to receive the trained model sent by the server, and is used for updating the fiber fatigue strength prediction model and fiber lifetime prediction model in the fiber lifetime prediction device.

[0147] Specifically, in this type of embodiment, the fiber lifetime prediction device cannot train or update the model on its own and must passively receive the already trained model.

[0148] In one possible implementation, the fiber lifetime prediction device further includes:

[0149] The acquisition unit is used to acquire multiple fatigue sample datasets and multiple performance parameters corresponding to the multiple fatigue sample datasets, as well as multiple remaining life values ​​corresponding to multiple prediction datasets.

[0150] The training unit is trained based on multiple fatigue sample datasets and multiple performance parameters corresponding to the multiple fatigue sample datasets to obtain an optical fiber fatigue strength prediction model; and trained based on multiple prediction datasets and multiple remaining lifetime values ​​corresponding to the multiple prediction datasets to obtain an optical fiber lifetime prediction model.

[0151] Specifically, in this optional embodiment, the fiber lifetime prediction device can receive samples and train or update the model based on the samples.

[0152] For further details, please see Figure 5 The diagram shown is a structural schematic of a fiber lifetime prediction training device provided in an embodiment of the present invention.

[0153] Specifically, the fiber lifetime prediction training device can be the aforementioned server or devices within a server, including an acquisition unit, a training unit, and a transmission unit, wherein:

[0154] The acquisition unit is used to acquire multiple fatigue sample datasets and multiple performance parameters corresponding to the multiple fatigue sample datasets, as well as multiple remaining life values ​​corresponding to multiple prediction datasets.

[0155] The training unit is trained based on multiple fatigue sample datasets and multiple performance parameters corresponding to the multiple fatigue sample datasets to obtain an optical fiber fatigue strength prediction model; and trained based on multiple prediction datasets and multiple remaining lifetime values ​​corresponding to the multiple prediction datasets to obtain an optical fiber lifetime prediction model.

[0156] The sending unit is used to send the trained model to the terminal for updating or replacing the model in the terminal.

[0157] For further details, please see Figure 6 The diagram shown is an interactive schematic of a fiber optic lifetime terminal and server according to an embodiment of the present invention.

[0158] Specifically, the interaction process and architecture between the terminal and the server provided in this embodiment of the invention are based on Figure 3 Implementation, including but not limited to the following steps:

[0159] Step S301: The terminal sends the original dataset and the corresponding touch judgment result to the server.

[0160] The terminal sends the original dataset and the corresponding remaining lifetime value to the server.

[0161] The server receives the original dataset and the corresponding remaining lifetime value, and inputs the original dataset and the remaining lifetime value into the fiber lifetime prediction training device for training; then the server sends the fiber lifetime prediction device to the terminal.

[0162] The terminal receives and updates the fiber lifetime prediction device, thereby completing the interaction between the two.

[0163] The terminal receives a fiber optic lifetime prediction device from the server, which is used to replace the original fiber optic lifetime prediction device.

[0164] In this embodiment of the invention, the terminal itself does not have the ability to perform model training. Therefore, the original dataset and the remaining lifetime value corresponding to the original dataset are sent to the server for model training, so as to update or replace the fiber lifetime prediction device in the terminal.

[0165] This invention provides a computer-readable storage medium storing a computer program, the computer program including program instructions, which, when executed by a processor, cause the processor to perform... Figure 3 The operations performed by the terminal in the above embodiment.

[0166] This invention also provides a computer program product that, when run on a processor, implements... Figure 3 The operations performed by the terminal in the above embodiment.

[0167] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0168] The present invention has been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of the invention. The descriptions of the embodiments above are only for the purpose of helping to understand the invention and its core ideas. It should be noted that those skilled in the art can make various improvements and modifications to the invention without departing from its principles, and these improvements and modifications also fall within the protection scope of the claims of the present invention.

Claims

1. A method for predicting the lifetime of an optical fiber, characterized in that, include: Using the service life, environmental parameters, and test data as input, the raw dataset of the optical fiber under test is formed; The original dataset is input into the optical fiber fatigue strength prediction model to obtain performance parameters; the performance parameters include at least fatigue parameters and Welb distribution slope. Using the factory parameters and the aforementioned performance parameters as data inputs, a prediction dataset for the optical fiber under test is formed. The prediction dataset is then input into the optical fiber lifetime prediction model to obtain the remaining lifetime value of the optical fiber under test. The fiber fatigue strength prediction model is a model trained based on the performance parameters corresponding to multiple fatigue sample datasets; the service life, environmental parameters and test data included in any of the fatigue sample datasets correspond to the fatigue parameters and Welb distribution slope included in the performance parameters. The fiber optic lifetime prediction model is a model trained based on the remaining lifetime values ​​corresponding to multiple prediction datasets. The factory parameters and performance parameters included in any of the prediction datasets correspond to the remaining lifetime value.

2. The method for predicting optical fiber lifetime as described in claim 1, characterized in that, The fatigue parameters include the following calculation steps: S10. Divide the optical fiber to be tested into multiple optical fiber samples with a length of S and a number of segments of N. S11. Pre-set constant strain rate, temperature, and relative humidity during the test; S12. Record the breaking tension T of all optical fiber samples from the 1st to the Nth fiber. S13. Calculate the fracture stress of all fiber optic samples: σ f =T / A g ;σ f The tensile stress is represented by T, and the tensile force is represented by Ag, which is the nominal cross-sectional area of ​​the glass optical fiber. S14. Calculate the stress rate of all the optical fibers under test: Represents the stress rate, t(σ) f ) represents the fracture time, t(0.8σ f This indicates the time taken to reach 80% of the fracture stress; S15. Calculate the fatigue parameters of the optical fiber under test: σ f (0.5) represents the fracture stress when the cumulative failure probability is 0.

5. nd Indicates fatigue parameters, σ1 This represents the fracture stress at a stress rate of 1.

3. The method for predicting optical fiber lifetime as described in claim 2, characterized in that, The slope of the Welb distribution is calculated using the following steps: S20. Rearrange all the optical fibers to be tested in order of their fracture stress, from lowest to highest as 1, 2, ..., N; S21. Calculate the cumulative breakage probability of each optical fiber under test: F i = (i-0.5) / N; F i This represents the cumulative fracture probability, where i takes values ​​of 1, 2, ..., N; S22. Plot ln[-ln(1-F)] i )] for ln(σ f The Weibull curve is plotted, and data is marked on the curve. S23. Based on the dynamic Weibull curve and cumulative function Calculate the Welb slope: m d =2.46 / ln[σ f (0.85)]-ln[σ f (0.15)];m d Let σ0 represent the slope of the Weibull distribution, and σ0 represent the Weibull parameter. f (0.85) represents the fracture stress when the cumulative failure probability is 0.85, σ f (0.15) represents the fracture stress when the cumulative failure probability is 0.

15.

4. The method for predicting optical fiber lifetime as described in claim 3, characterized in that, The remaining lifetime value includes the following calculation steps: S30. Based on the predicted dataset, calculate the remaining lifetime of the optical fiber under test: t f This indicates the remaining lifetime of the optical fiber under test. σ represents the adjustment factor. p Indicates screening stress, t p Indicates the effective screening time, F represents the fiber failure rate, and N represents the effective screening time. p σ represents the average breakage rate per unit length obtained from the fiber optic factory screening test. a The static stress on the optical fiber is represented by n, which represents the actual fatigue parameter of the optical fiber and reflects the microcrack propagation rate. L represents the equivalent tensile length.

5. The method for predicting optical fiber lifetime as described in claim 4, characterized in that, The service life is t i ; The factory parameters include σ p t p F, N p σ a N p The specific values ​​of L and L are provided by the supplier of the optical fiber to be tested. The environmental parameters include atmospheric pressure, maximum operating temperature, minimum operating temperature, operating temperature difference, maximum operating wind speed level, minimum operating wind speed level, average operating wind speed level, and number of lightning strikes. The number of lightning strikes is an independent value, denoted as Y; the characteristic values ​​of the environmental parameters other than the number of lightning strikes are denoted as X. j j = 1, 2, ..., 7.

6. The method for predicting optical fiber lifetime as described in claim 4, characterized in that, The performance parameters include median breakage time and maximum stress on the optical fiber; The formula for calculating the median fracture time is as follows: The formula for calculating the maximum stress on the optical fiber is: σ=E0(1+cε)ε, where σ represents the stress on the optical fiber, E0 represents Young's modulus, c represents the material correlation coefficient, and ε represents the maximum strain of the optical fiber.

7. The method for predicting optical fiber lifetime as described in claim 1, characterized in that, The steps for constructing the optical fiber fatigue strength prediction model and / or the optical fiber lifetime prediction model are as follows: S40. Feature extraction is performed using a convolutional neural network; S41. Learn the feature relationships on the time series through a long short-term memory network; S42. Form a model network.

8. An apparatus for predicting the lifetime of an optical fiber, used to implement the method for predicting the lifetime of an optical fiber as described in any one of claims 1-7, characterized in that, include: The generation unit is used to generate the raw dataset based on the service life, environmental parameters, and test data of the optical fiber under test. Used to generate a prediction dataset based on the factory parameters and performance parameters of the optical fiber under test; The processing unit is used to input the original dataset into the optical fiber fatigue strength prediction model to obtain performance parameters; This is used to input the predicted dataset into the optical fiber lifetime prediction model and output the remaining lifetime value of the optical fiber under test.

9. A terminal, characterized in that, The terminal includes at least one processor, a communication interface, and a memory. The communication interface is used to send and / or receive data, the memory is used to store computer programs, and the at least one processor is used to call the computer programs stored in at least one memory to implement a fiber optic lifetime prediction method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when run on a processor, implements a method for predicting the lifetime of an optical fiber as described in any one of claims 1-7.

Images