A Machine Learning-Based Method for Fiber Optic Fault Prediction and Early Warning

Abstract

This invention discloses a machine learning-based method for optical fiber fault prediction and early warning. The method includes: acquiring multi-dimensional optical feature data of the optical fiber; labeling fault areas to obtain labeled feature data; preprocessing the labeled feature data to construct a standard optical fiber fault feature dataset; inputting time-series optical feature data into a neural network model to obtain labeled feature data for each time period and the future, and correcting it using the standard dataset; inputting real-time data of the optical fiber under test into the model, and outputting fault data and predicted data. This method uses machine learning to process multi-dimensional features, achieving accurate prediction and early warning of optical fiber faults, improving the timeliness and accuracy of fault detection.

Description

Technical Field

[0001] This invention relates to the field of technology, specifically to a method for optical fiber fault prediction and early warning based on machine learning. Background Technology

[0002] In fiber optic communication systems, optical fibers serve as the core carrier of information transmission, and the stability of their operation directly determines the communication quality. However, during long-term use, optical fibers are susceptible to various faults caused by external environmental factors, physical damage, and aging, such as fiber breakage and abnormal attenuation. These faults can lead to communication interruptions or signal quality degradation, causing significant inconvenience to production and daily life.

[0003] Currently, fiber optic fault detection mainly relies on manual inspections and traditional monitoring methods. Manual inspections are not only time-consuming and labor-intensive, but also difficult to monitor the fiber's condition in real time, often only discovering problems after they occur, making early prediction and warning impossible. Traditional monitoring methods typically only monitor single characteristic parameters, such as optical power and attenuation coefficient, making it difficult to comprehensively consider the multidimensional characteristics of the fiber, resulting in low accuracy in fault identification and an inability to effectively predict fault development trends.

[0004] With the development of machine learning technology, it has shown great potential in the field of fault prediction and diagnosis. Applying machine learning technology to fiber optic fault prediction and early warning can comprehensively process multi-dimensional feature data of optical fibers, improving the accuracy of fault identification and the timeliness of prediction. Therefore, developing a machine learning-based method for fiber optic fault prediction and early warning has significant practical implications. Summary of the Invention

[0005] To address the aforementioned problems, this invention provides a machine learning-based method for optical fiber fault prediction and early warning. By fusing temporal and spatial features and dynamically correcting the model, it achieves early prediction and accurate early warning of faults.

[0006] The objective of this invention is achieved through the following technical solution:

[0007] A machine learning-based method for optical fiber fault prediction and early warning includes the following steps:

[0008] S1. Obtain multidimensional optical characteristic data of the optical fiber;

[0009] S2. Mark the fault areas of the multidimensional optical feature data to obtain marked feature data;

[0010] S3. Preprocess the labeled feature data to construct a standard optical fiber fault feature dataset;

[0011] S4. Input the temporal optical feature data in the multidimensional optical feature data into the neural network model to obtain the labeled feature data of each time period in the temporal optical features and the predicted labeled feature data of the future time period, and correct it with the standard optical fiber fault feature dataset.

[0012] S5. Input the multi-dimensional optical feature data of the optical fiber under test into the neural network model of S4 in real time, and output the optical fiber fault data and prediction data.

[0013] As a preferred embodiment, the multidimensional optical feature data includes at least deformation features and transmission features;

[0014] The deformation characteristics are determined based on the surface curvature of the optical fiber, including the following steps:

[0015] Obtain the surface curvature of the fiber surface and its internal points, and determine whether the surface curvature at each point is within a preset threshold range;

[0016] If the surface curvature of the current point is not within the preset threshold range, obtain the surface curvature of the adjacent points of the current point until the surface curvature of a point is within the preset threshold range.

[0017] The deformation feature is obtained by acquiring multiple deformation anomalies formed by the set of points.

[0018] The transmission characteristics are determined based on the transmission parameters of the optical fiber, including the following steps:

[0019] Acquire the optical signal transmission signal from the deformed feature portion to another deformed feature portion or non-deformed feature portion, and determine the current optical attenuation coefficient;

[0020] A sequence of optical attenuation coefficients is obtained, and each unit of the sequence is set as the optical attenuation coefficient and transmission distance from the current deformed feature to a deformed feature or a non-deformed feature.

[0021] A set of optical decay coefficients is obtained, which is configured to include a sequence of optical decay coefficients for each of the deformed features;

[0022] Transmission characteristics are obtained by comparing each unit of the optical attenuation coefficient set with the preset optical attenuation coefficient threshold range in the transmission parameters.

[0023] As a preferred method, the fault area is marked in step S2, which includes the following steps:

[0024] When the surface curvature deviation of the abnormal deformation part exceeds the preset threshold and the corresponding optical decay coefficient exceeds the threshold range but does not fluctuate significantly, it is marked as the first fault type.

[0025] When the difference in optical decay coefficient between adjacent units in the optical decay coefficient sequence exceeds the preset fluctuation threshold, and the surface curvature deviation of the abnormal deformation part is within the allowable range, it is marked as the second fault type.

[0026] When the optical attenuation coefficients corresponding to the abnormal deformation exceeding the preset proportion all exceed the threshold range, and the optical attenuation coefficients corresponding to the transmission spacing are generally high, it is marked as the third fault type.

[0027] As a preferred embodiment, the preprocessing includes:

[0028] Remove coordinate points in the deformation characteristics that exceed the physical size range of the optical fiber, and abnormal optical attenuation coefficients in the transmission characteristics that are outside the distortion range;

[0029] The surface curvature value in the deformation characteristics is normalized to the [0,1] interval based on the curvature threshold of the standard optical fiber, and the optical attenuation coefficient in the transmission characteristics is normalized based on the theoretical optimal attenuation rate.

[0030] Establish a mapping relationship between the coordinates of the deformation anomaly and the corresponding optical decay coefficient sequence to form an associated feature vector.

[0031] As a preferred embodiment, the neural network model in step S4 includes:

[0032] The temporal feature extraction layer employs a long short-term memory network to extract temporal correlation features from temporal optical feature data, which includes the sequential changes in optical decay coefficients over continuous time periods and the temporal data of the location extension of deformation anomalies.

[0033] The spatial feature extraction layer uses a convolutional neural network to extract the spatial distribution features of the three-dimensional coordinates of the deformation anomaly in the deformation features.

[0034] The fusion prediction layer combines the outputs of the temporal feature extraction layer and the spatial feature extraction layer, outputting labeled feature data for each time period and predicted labeled feature data for future time periods.

[0035] As a preferred approach, step S4 involves correction using a standard fiber optic fault feature dataset, including the following steps:

[0036] The predicted labeled feature data is compared with the features of the same type of fault in the standard fiber optic fault feature dataset, and the feature deviation value is calculated.

[0037] If the deviation exceeds the preset correction threshold, the weight parameters of the neural network model are adjusted through the backpropagation algorithm, and the correction is repeated until the deviation is within the preset allowable range.

[0038] As a preferred approach, when outputting prediction data, the influence of the expansion of the simulated deformation region on the transmission coefficient is calculated, the three-dimensional coordinate expansion range of the deformation anomaly and the corresponding change in the transmission coefficient are predicted within a preset time period, and the prediction result of the fault development trend is output.

[0039] The beneficial effects of this invention are as follows:

[0040] This invention acquires multidimensional optical characteristic data of optical fibers, including deformation and transmission characteristics, which can comprehensively reflect the state of optical fibers and provide rich evidence for fault prediction and early warning.

[0041] Accurate labeling of fault areas and classification of different fault types can help improve the accuracy of neural network models in identifying faults.

[0042] Preprocessing labeled feature data, removing outliers, normalizing the data, and establishing associated feature vectors can improve the quality and consistency of the data, laying a good foundation for model training and prediction.

[0043] By employing a neural network model that includes a temporal feature extraction layer, a spatial feature extraction layer, and a fusion prediction layer, temporal and spatial features can be effectively integrated to improve the ability to predict faults.

[0044] The prediction results were corrected by using a standard fiber optic fault feature dataset, and the model parameters were continuously optimized, which further improved the accuracy of the prediction.

[0045] It can output predictions of fault development trends, enabling early prediction and warning of fiber optic faults, facilitating timely action by staff to minimize losses caused by the faults. Attached Figure Description

[0046] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0047] Figure 1 This is a flowchart of a method according to an embodiment of the present invention. Detailed Implementation

[0048] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0049] This disclosure provides a machine learning-based method for optical fiber fault prediction and early warning, such as... Figure 1 As shown, it includes the following steps:

[0050] S1. Obtain multidimensional optical characteristic data of the optical fiber;

[0051] S2. Mark the fault areas of the multidimensional optical feature data to obtain marked feature data;

[0052] S3. Preprocess the labeled feature data to construct a standard optical fiber fault feature dataset;

[0053] S4. Input the temporal optical feature data in the multidimensional optical feature data into the neural network model to obtain the labeled feature data of each time period in the temporal optical features and the predicted labeled feature data of the future time period, and correct it with the standard optical fiber fault feature dataset.

[0054] S5. Input the multi-dimensional optical feature data of the optical fiber under test into the neural network model of S4 in real time, and output the optical fiber fault data and prediction data.

[0055] As a preferred embodiment, the multidimensional optical feature data includes at least deformation features and transmission features;

[0056] The deformation characteristics are determined based on the surface curvature of the optical fiber, including the following steps:

[0057] Obtain the surface curvature of the fiber surface and its internal points, and determine whether the surface curvature at each point is within a preset threshold range;

[0058] If the surface curvature of the current point is not within the preset threshold range, obtain the surface curvature of the adjacent points of the current point until the surface curvature of a point is within the preset threshold range.

[0059] The deformation feature is obtained by acquiring multiple deformation anomalies formed by the set of points.

[0060] The transmission characteristics are determined based on the transmission parameters of the optical fiber, including the following steps:

[0061] Acquire the optical signal transmission signal from the deformed feature portion to another deformed feature portion or non-deformed feature portion, and determine the current optical attenuation coefficient;

[0062] A sequence of optical attenuation coefficients is obtained, and each unit of the sequence is set as the optical attenuation coefficient and transmission distance from the current deformed feature to a deformed feature or a non-deformed feature.

[0063] A set of optical decay coefficients is obtained, which is configured to include a sequence of optical decay coefficients for each of the deformed features;

[0064] Transmission characteristics are obtained by comparing each unit of the optical attenuation coefficient set with the preset optical attenuation coefficient threshold range in the transmission parameters.

[0065] As a preferred method, the fault area is marked in step S2, which includes the following steps:

[0066] When the surface curvature deviation of the abnormal deformation part exceeds the preset threshold and the corresponding optical decay coefficient exceeds the threshold range but does not fluctuate significantly, it is marked as the first fault type.

[0067] When the difference in optical decay coefficient between adjacent units in the optical decay coefficient sequence exceeds the preset fluctuation threshold, and the surface curvature deviation of the abnormal deformation part is within the allowable range, it is marked as the second fault type.

[0068] When the optical attenuation coefficients corresponding to the abnormal deformation exceeding the preset proportion all exceed the threshold range, and the optical attenuation coefficients corresponding to the transmission spacing are generally high, it is marked as the third fault type.

[0069] As a preferred embodiment, the preprocessing includes:

[0070] Remove coordinate points in the deformation characteristics that exceed the physical size range of the optical fiber, and abnormal optical attenuation coefficients in the transmission characteristics that are outside the distortion range;

[0071] The surface curvature value in the deformation characteristics is normalized to the [0,1] interval based on the curvature threshold of the standard optical fiber, and the optical attenuation coefficient in the transmission characteristics is normalized based on the theoretical optimal attenuation rate.

[0072] Establish a mapping relationship between the coordinates of the deformation anomaly and the corresponding optical decay coefficient sequence to form an associated feature vector.

[0073] As a preferred embodiment, the neural network model in step S4 includes:

[0074] The temporal feature extraction layer employs a long short-term memory network to extract temporal correlation features from temporal optical feature data, which includes the sequential changes in optical decay coefficients over continuous time periods and the temporal data of the location extension of deformation anomalies.

[0075] The spatial feature extraction layer uses a convolutional neural network to extract the spatial distribution features of the three-dimensional coordinates of the deformation anomaly in the deformation features.

[0076] The fusion prediction layer combines the outputs of the temporal feature extraction layer and the spatial feature extraction layer, outputting labeled feature data for each time period and predicted labeled feature data for future time periods.

[0077] As a preferred approach, step S4 involves correction using a standard fiber optic fault feature dataset, including the following steps:

[0078] The predicted labeled feature data is compared with the features of the same type of fault in the standard fiber optic fault feature dataset, and the feature deviation value is calculated.

[0079] If the deviation exceeds the preset correction threshold, the weight parameters of the neural network model are adjusted through the backpropagation algorithm, and the correction is repeated until the deviation is within the preset allowable range.

[0080] As a preferred approach, when outputting prediction data, the influence of the expansion of the simulated deformation region on the transmission coefficient is calculated, the three-dimensional coordinate expansion range of the deformation anomaly and the corresponding change in the transmission coefficient are predicted within a preset time period, and the prediction result of the fault development trend is output.

[0081] Specifically, S1, acquire multidimensional optical characteristic data of the optical fiber, which includes deformation characteristics and transmission characteristics.

[0082] When acquiring deformation characteristics, the surface curvature of each point on and inside the optical fiber is obtained using relevant detection equipment, with a preset surface curvature threshold range of [R1, R2]. For each point, it is determined whether its surface curvature is within [R1, R2]. If not, the surface curvature of the adjacent points is obtained until a point with surface curvature within [R1, R2] is found. The set of these points is taken as the deformation anomaly, and multiple deformation anomalies together constitute the deformation characteristics.

[0083] When acquiring transmission characteristics, the optical signal transmission signals from each deformable feature to other deformable or non-deformable feature are first acquired, and the current optical attenuation coefficient is calculated based on the optical signal transmission signals. Then, an optical attenuation coefficient sequence is constructed, where each unit contains the optical attenuation coefficient and transmission distance from the current deformable feature to another deformable or non-deformable feature. The optical attenuation coefficient sequences of all deformable feature parts are combined into an optical attenuation coefficient set, and each unit in the optical attenuation coefficient set is compared with a preset optical attenuation coefficient threshold interval [K1, K2] to obtain the transmission characteristics.

[0084] S2. Mark the fault area of ​​the multidimensional optical feature data. The preset threshold for the surface curvature deviation of the abnormal deformation part is ΔR, the threshold range of the optical decay coefficient is [K1,K2], the preset fluctuation threshold for the optical decay coefficient difference is ΔK, and the preset ratio is 50%.

[0085] When the surface curvature deviation of the abnormal deformation part exceeds ΔR, and the corresponding optical decay coefficient exceeds [K1,K2] but does not fluctuate significantly, it is marked as the first fault type;

[0086] When the difference in optical decay coefficient between adjacent units in the optical decay coefficient sequence exceeds ΔK, and the surface curvature deviation of the deformed abnormal part is within the allowable range, it is marked as the second fault type.

[0087] When the optical attenuation coefficients corresponding to more than 50% of the abnormal deformation parts all exceed [K1,K2], and the optical attenuation coefficients corresponding to the transmission spacing are generally high, it is marked as the third type of fault.

[0088] S3. Preprocess the labeled feature data to construct a standard fiber optic fault feature dataset.

[0089] The preprocessing process is as follows:

[0090] Remove coordinate points in the deformation characteristics that exceed the physical size range of the optical fiber. For example, if the actual length of the optical fiber is L, remove coordinate points whose length direction exceeds L. Remove abnormal optical attenuation coefficients in the transmission characteristics that are outside the distortion range, such as optical attenuation coefficients that are negative or much larger than the normal range.

[0091] The surface curvature value in the deformation characteristics is normalized based on the curvature threshold R0 of the standard optical fiber. The normalization formula is: Normalized curvature = (Actual curvature - R0) / (Rmax - Rmin), where Rmax and Rmin are the preset maximum and minimum curvature values, respectively, so that the normalized curvature is in the range [0,1]. The optical attenuation coefficient in the transmission characteristics is normalized based on the theoretical optimal attenuation rate K0. The normalized optical attenuation coefficient = (Actual optical attenuation coefficient - K0) / (Kmax - Kmin), where Kmax and Kmin are the preset maximum and minimum optical attenuation coefficient values, respectively.

[0092] Establish a mapping relationship between the three-dimensional coordinates of the deformation anomaly and the corresponding optical decay coefficient sequence to form an associated feature vector. For example, for the deformation anomaly A (x1, y1, z1), its corresponding optical decay coefficient sequence is [K11, d11; K12, d12; ...; K1n, d1n], then the associated feature vector is (x1, y1, z1, K11, d11, K12, d12, ..., K1n, d1n).

[0093] S4. Input the temporal optical feature data into the neural network model and make corrections.

[0094] The neural network model includes a temporal feature extraction layer, a spatial feature extraction layer, and a fusion prediction layer.

[0095] The temporal feature extraction layer uses a Long Short-Term Memory (LSTM) network. It takes the changes in the light decay coefficient sequence and the location of the deformation anomaly within a continuous time period as input to expand the temporal data. For example, data is collected once per hour, and 100 hours of data are collected as input to extract the temporal correlation features.

[0096] The spatial feature extraction layer uses a convolutional neural network (CNN) to extract the spatial distribution features of the three-dimensional coordinates of the deformation anomaly in the deformation features, and obtains spatial features through convolution and pooling operations.

[0097] The fusion prediction layer combines the outputs of the temporal feature extraction layer and the spatial feature extraction layer. For example, it uses a fully connected layer to concatenate and process the two types of features, outputting labeled feature data for each time period and predicted labeled feature data for future time periods.

[0098] Correction process: The predicted labeled feature data is compared with the features of the same type of fault in the standard fiber optic fault feature dataset. The feature deviation value is calculated. For example, for the first fault type, the difference between the predicted surface curvature deviation and the surface curvature deviation in the standard data, and the difference between the predicted optical attenuation coefficient and the optical attenuation coefficient in the standard data are calculated, and the feature deviation value is obtained by combining them. The preset correction threshold is 0.1. If the deviation value exceeds 0.1, the weight parameters of the neural network model are adjusted through the backpropagation algorithm, and the training and correction process is repeated until the deviation value is within 0.1.

[0099] S5. Perform fault prediction and early warning for the optical fiber under test.

[0100] The optical fiber under test is sampled in real time to obtain its multidimensional optical feature data. After preprocessing according to the method in step S3, it is input into the neural network model trained in step S4.

[0101] The model outputs fiber optic fault data, indicating whether a fault exists in the fiber and its type. Simultaneously, it outputs prediction data, simulating the impact of deformation region expansion on the transmission coefficient. For example, assuming the deformation anomaly expands at a certain rate, it calculates the 3D coordinate expansion range at different time points and the corresponding transmission coefficient change based on the expansion range. It then predicts the 3D coordinate expansion range of the deformation anomaly and the corresponding transmission coefficient change within the next 24 hours, outputting the predicted fault development trend, such as, "Within the next 12 hours, the deformation anomaly of the first fault type will expand from (x1, y1, z1) to (x2, y2, z2), and the optical attenuation coefficient will increase from K1 to K2."

[0102] The foregoing description and accompanying drawings fully illustrate embodiments of this disclosure to enable those skilled in the art to practice them. Other embodiments may include structural, logical, electrical, procedural, and other changes. The embodiments represent only possible variations. Individual components and functions are optional unless explicitly required, and the order of operations may vary. Parts and features of some embodiments may be included in or replace parts and features of other embodiments. Moreover, the terminology used in this application is for describing embodiments only and is not intended to limit the claims. As used in the description of embodiments and claims, the singular forms “a,” “an,” and “the” are intended to equally include the plural forms unless the context clearly indicates otherwise. Similarly, the term “and / or” as used in this application means including one or more of the associated listed items and all possible combinations thereof. Additionally, when used in this application, the term "comprise" and its variations "comprises" and / or "comprising" refer to the presence of stated features, integrals, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or groups thereof. Without further limitations, an element defined by the phrase "comprises a..." does not exclude the presence of other identical elements in the process, method, or apparatus that includes said element. In this document, each embodiment may focus on the differences from other embodiments, and similar or identical parts between embodiments can be referred to mutually. For methods, products, etc., disclosed in the embodiments, if they correspond to the method section disclosed in the embodiments, the relevant parts can be referred to the description of the method section.

[0103] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented using electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods for each specific application to achieve the described functions, but such implementation should not be considered beyond the scope of the embodiments of this disclosure. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the described devices, apparatuses, and units can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0104] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, function, and operation of possible implementations of apparatus, methods, and computer program products according to embodiments of the present disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. In the descriptions corresponding to the flowcharts and block diagrams in the accompanying drawings, the operations or steps corresponding to different blocks may also occur in a different order than those disclosed in the description; sometimes there is no specific order between different operations or steps. For example, two consecutive operations or steps may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. Each block in a block diagram and / or flowchart, and combinations of blocks in a block diagram and / or flowchart, can be implemented using a dedicated hardware-based device that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

Claims

1. A machine learning based optical fiber failure prediction and early warning method, characterized in that, The method comprises the following steps: S1, obtaining multi-dimensional optical characteristic data of the optical fiber; S2, labeling a failure region of the multi-dimensional optical characteristic data to obtain labeled characteristic data; S3, preprocessing the labeled characteristic data to construct a standard optical fiber failure characteristic data set; S4, inputting time-series optical characteristic data in the multi-dimensional optical characteristic data into a neural network model to obtain labeled characteristic data of each time period in the time-series optical characteristic data and predicted labeled characteristic data of a future time period, and correcting the labeled characteristic data and the predicted labeled characteristic data by using the standard optical fiber failure characteristic data set; S5, inputting multi-dimensional optical characteristic data of a to-be-tested optical fiber sampled in real time into the neural network model in S4 to output optical fiber failure data and prediction data; The multi-dimensional optical characteristic data at least comprises deformation characteristics and transmission characteristics; The deformation characteristics are determined according to surface curvatures of the optical fiber, and the method comprises the following steps: obtaining surface curvatures of each point on the surface and inside the optical fiber, and judging whether the surface curvature of each point is within a preset threshold interval; if the surface curvature of the point is not within the preset threshold interval, obtaining the surface curvature of the position of a neighboring point of the current point until the surface curvature of a point is within the preset threshold interval; obtaining a plurality of deformation abnormal parts formed by a point set to obtain the deformation characteristics; The transmission characteristics are determined according to transmission parameters of the optical fiber, and the method comprises the following steps: obtaining an optical signal transmission signal from the deformation characteristic part to another deformation characteristic part or a non-deformation characteristic part, and determining a current optical attenuation coefficient; obtaining an optical attenuation coefficient sequence, each unit of the sequence being set as the optical attenuation coefficient and a transmission interval from the deformation characteristic part to one deformation characteristic part or non-deformation characteristic part; obtaining an optical attenuation coefficient set, the optical attenuation coefficient set being configured to comprise the optical attenuation coefficient sequence of each deformation characteristic part; comparing each unit of the optical attenuation coefficient set with a preset optical attenuation coefficient threshold interval in the transmission parameter to obtain the transmission characteristics; The step S2 of labeling the failure region comprises the following steps: when the surface curvature deviation of the deformation abnormal part exceeds the preset threshold and the corresponding optical attenuation coefficient exceeds the threshold interval but does not fluctuate significantly, the first failure type is labeled; when the optical attenuation coefficient difference of adjacent units in the optical attenuation coefficient sequence exceeds a preset fluctuation threshold, and the surface curvature deviation of the deformation abnormal part is within the allowable range, the second failure type is labeled; when the optical attenuation coefficient corresponding to more than a preset proportion of deformation abnormal parts exceeds the threshold interval, and the optical attenuation coefficient corresponding to the transmission interval is generally high, the third failure type is labeled. 2.The machine learning based optical fiber failure prediction and warning method according to claim 1, characterized in that, The preprocessing comprises: eliminating coordinate points in the deformation characteristics that exceed the physical size range of the optical fiber, and eliminating abnormal optical attenuation coefficients outside the distortion range in the transmission characteristics; normalizing the surface curvature value in the deformation characteristics to the [0, 1] interval based on the curvature threshold of the standard optical fiber, and normalizing the optical attenuation coefficient in the transmission characteristics based on the theoretical optimal attenuation rate; establishing a mapping relationship between the coordinates of the deformation abnormal part and the corresponding optical attenuation coefficient sequence to form an associated characteristic vector. 3.The machine learning based optical fiber failure prediction and warning method according to claim 2, characterized in that, The neural network model in the step S4 comprises: The time sequence feature extraction layer adopts a long short-term memory network to extract time correlation features in time sequence optical feature data, the time sequence optical feature data including a sequence change of an optical attenuation coefficient in a continuous time period and position expansion time sequence data of a deformation abnormal part; The spatial feature extraction layer adopts a convolutional neural network to extract spatial distribution features of three-dimensional coordinates of the deformation abnormal part in the deformation feature; The fusion prediction layer fuses outputs of the time sequence feature extraction layer and the spatial feature extraction layer to output labeled feature data of each time period and predicted labeled feature data of a future time period.

4. The machine learning based optical fiber failure prediction and warning method according to claim 3, characterized in that, The step S4 of correcting the standard optical fiber fault feature data set includes the following steps: The predicted labeled feature data is compared with features of the same type of fault in the standard optical fiber fault feature data set to calculate a feature deviation value; If the deviation value exceeds a preset correction threshold, the weight parameters of the neural network model are adjusted through a back propagation algorithm, and the correction is repeated until the deviation value is within a preset allowable range.

5. The machine learning based optical fiber failure prediction and warning method according to claim 4, characterized in that, When the prediction data is output, the influence of the expansion of the deformation area on the transmission coefficient is simulated, the three-dimensional coordinate expansion range of the deformation abnormal part and the corresponding transmission coefficient change value in a future preset time period are predicted, and a prediction result of the fault development trend is output.

Images