CN-122001460-A - Optical cable fault identification method and device, electronic equipment and storage medium

CN122001460ACN 122001460 ACN122001460 ACN 122001460ACN-122001460-A

Abstract

The application provides an optical cable fault identification method, which comprises the steps of preprocessing optical cable monitoring data acquired in advance to obtain training data and test data, training a neural network model according to the training data to obtain an optical cable fault identification model, wherein the optical cable fault identification model comprises a feature extraction layer, a regional attention module and a residual feature fusion structure, carrying out parameter adjustment on the optical cable fault identification model according to the test data to obtain an optimized fault identification model, and inputting monitoring data to be identified into the optimized fault identification model to obtain a fault identification result. Therefore, intelligent analysis and positioning capability of digital optical cable faults in a complex power communication network is improved, and safe and efficient operation of optical fiber communication is ensured.

Inventors

LIU PENG
DING JIANBO
ZHANG XINQIAO
LIU XU
LIU DI
ZHANG SHANHUI
LI LONGMEI

Assignees

北京中电飞华通信有限公司

Dates

Publication Date: 20260508
Application Date: 20260130
Priority Date: 20251208

Claims (10)

1. A method of identifying a fiber optic cable fault, comprising: Preprocessing the pre-acquired optical cable monitoring data to obtain training data and test data; Training the neural network model according to the training data to obtain an optical cable fault recognition model, wherein the optical cable fault recognition model comprises a feature extraction layer, a regional attention module and a residual feature fusion structure; Performing parameter adjustment on the optical cable fault recognition model according to the test data to obtain an optimized fault recognition model; And inputting the monitoring data to be identified into the optimized fault identification model to obtain a fault identification result.
2. The method for identifying a cable fault according to claim 1, wherein the preprocessing the pre-acquired cable monitoring data to obtain training data and test data specifically comprises: Dividing the optical cable monitoring data into a plurality of sample data according to a preset region segment, and marking the fault type of the sample data to obtain marked fault data, wherein the optical cable monitoring data comprises an OTDR echo curve, a reflection intensity sequence and a weak event waveform; and converting the marked fault data into a two-dimensional feature map to obtain the training data and the test data.
3. The method for identifying a cable failure according to claim 1, wherein training the neural network model according to the training data to obtain a cable failure identification model specifically comprises: adjusting the training data according to a preset data enhancement algorithm to obtain training enhancement data; inputting the training enhancement data into the neural network model so that the neural network model outputs a training recognition result; And optimizing the neural network model according to a preset loss function and the training recognition result to obtain the optical cable fault recognition model.
4. The method for identifying a fiber optic cable fault of claim 3, wherein the inputting the training enhancement data into the neural network model to cause the neural network model to output a training identification result comprises: Inputting the training enhancement data to the feature extraction layer such that the feature extraction layer outputs a feature depth representation; inputting the characteristic depth representation to the region attention module so that the region attention module marks a fault region in the training data according to the characteristic depth representation to obtain marked training data; And inputting the labeling training data into the residual characteristic fusion structure so that the residual characteristic fusion structure outputs the training recognition result.
5. The method for identifying a cable failure according to claim 1, wherein the performing parameter adjustment on the cable failure identification model according to the test data to obtain an optimized failure identification model specifically comprises: inputting the test model into the optical cable fault recognition model to obtain a test recognition result; And carrying out parameter adjustment on the optical cable fault recognition model according to the test recognition result to obtain the optimized fault recognition model.
6. The fiber optic cable fault identification method of claim 1, wherein after said inputting the monitoring data to be identified to the optimized fault identification model to obtain a fault identification result, the method further comprises: performing model performance evaluation according to actual fault information and the fault identification result to obtain model performance indexes, wherein the model performance indexes comprise fault identification accuracy, weak event detection rate, positioning average error and confidence degree distribution; And carrying out iterative optimization on the optimized fault recognition model according to the model performance index.
7. An optical cable fault identification device comprising: The pretreatment module is used for carrying out pretreatment on the optical cable monitoring data obtained in advance to obtain training data and test data; The training module is used for training the neural network model according to the training data to obtain an optical cable fault recognition model; The optimization module is used for carrying out parameter adjustment on the optical cable fault recognition model according to the test data to obtain an optimized fault recognition model; And the result output module is used for inputting the monitoring data to be identified into the optimized fault identification model to obtain a fault identification result.
8. The fiber optic cable fault identification device of claim 7 wherein said preprocessing module specifically comprises: The optical cable monitoring system comprises an area dividing module, a data processing module and a data processing module, wherein the area dividing module is used for dividing the optical cable monitoring data into a plurality of sample data according to a preset area segment, and marking fault types of the sample data to obtain marked fault data; and the data conversion module is used for converting the marked fault data into a two-dimensional feature map to obtain the training data and the test data.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the method of any one of claims 1 to 6 when the program is executed.
10. A non-transitory computer readable storage medium storing computer instructions for causing a computer to perform the method of any one of claims 1 to 6.

Description

Optical cable fault identification method and device, electronic equipment and storage medium Technical Field The application relates to the technical field of power communication, in particular to an optical cable fault identification method. Background In a modern power backbone communication network, along with the construction promotion of an extra-high voltage power grid and an energy internet, the optical fiber communication scale is rapidly enlarged, the optical cable structure is increasingly complex, and the network topology presents multi-level characteristics of cross sites and cross regions, so that the fault identification and positioning face higher accuracy and real-time requirements. However, under the transmission conditions of long distance and multiple interferences, the amplitude, phase and polarization characteristics of the optical signal are easy to be disturbed, and the problems of fuzzy characteristics, difficult identification of weak faults, sensitive noise and the like often occur in the existing detection technology, so that the requirements of a future power communication system on high reliability and high autonomous capacity are difficult to be met. Therefore, an intelligent fault detection method with higher robustness and stronger feature extraction capability is needed. In recent years, the deep learning technology has obvious advantages in the field of complex signal processing, can automatically extract stable features from large-scale data, and effectively solves the problems of noise disturbance and signal degradation which are difficult to process by a traditional algorithm. The capability provides a new implementation way for improving the recognition precision of the fault mode of the optical cable and enhancing the weak feature capturing capability. However, the digital optical cable fault detection at present still mainly depends on the traditional optical time domain reflectometer (Optical Time Domain Reflectometer, OTDR) analysis, template matching or shallow machine learning model, and is difficult to adapt to complex loss modes and multi-type fault scenes in a large-scale power communication network. The traditional method has limited resolving power to the weak characteristics covered by noise, and the missed detection, false detection and characteristic distortion are easy to occur in long-distance links. Meanwhile, the conventional convolution network has insufficient attention to a fault area, and is difficult to effectively highlight a key abnormal area under a multi-interference background, so that the positioning accuracy and stability are insufficient. Thus, the prior art still has significant limitations in terms of feature depth, region sensitivity, and complex scene robustness. Disclosure of Invention In view of the foregoing, the present application aims to provide a method, a device, an electronic device and a storage medium for identifying a fault of an optical cable. Based on the above purpose, the application provides an optical cable fault identification method, which comprises the steps of preprocessing optical cable monitoring data acquired in advance to obtain training data and test data. And training the neural network model according to the training data to obtain the optical cable fault recognition model. The optical cable fault recognition model comprises a feature extraction layer, a regional attention module and a residual feature fusion structure. And carrying out parameter adjustment on the optical cable fault recognition model according to the test data to obtain an optimized fault recognition model. And inputting the monitoring data to be identified into an optimized fault identification model to obtain a fault identification result. In some embodiments, pre-processing the pre-acquired optical cable monitoring data to obtain training data and test data specifically comprises dividing the optical cable monitoring data into a plurality of sample data according to a preset area segment, and marking fault types of the sample data to obtain marked fault data. The optical cable monitoring data comprise an OTDR echo curve, a reflection intensity sequence and a weak event waveform. And converting the marked fault data into a two-dimensional feature map to obtain training data and test data. In some embodiments, training the neural network model according to the training data to obtain the optical cable fault recognition model specifically comprises the steps of adjusting the training data according to a preset data enhancement algorithm to obtain training enhancement data. And inputting the training enhancement data into the neural network model so that the neural network model outputs training recognition results. And optimizing the neural network model according to a preset loss function and a training recognition result to obtain an optical cable fault recognition model. In some embodiments, the cable fault identification model include