An operation and maintenance intelligent terminal based on data analysis

Abstract

The present invention discloses an operation and maintenance intelligent terminal based on data analysis, which relates to the technical field of operation and maintenance intelligent terminals, and includes an initial operation and maintenance cycle setting module, a real-time data collection and monitoring module, a data preprocessing and feature extraction module, an intelligent evaluation and prediction module, and a dynamic maintenance strategy adjustment module; the initial operation and maintenance cycle setting module sets an initial operation and maintenance cycle for the fiber optic switch, and performs routine maintenance, inspection and optimization according to the established cycle after the fiber optic switch runs. The operation and maintenance intelligent terminal of the present invention intelligently identifies potential hidden dangers of link anomalies by collecting fiber optic switch data in real time, combining feature engineering and machine learning models, dynamically adjusts the maintenance cycle according to the severity of the anomalies, improves the fault warning ability and maintenance efficiency, avoids the lag of the fixed cycle, ensures the stable operation of the fiber optic link, reduces network interruptions or service failures caused by failure to detect problems in time, and thus improves the reliability and operation efficiency of the network.

Description

Technical Field

[0001] The present invention relates to the technical field of operation and maintenance intelligent terminals, and particularly to an operation and maintenance intelligent terminal based on data analysis. Background Art

[0002] An operation and maintenance intelligent terminal based on data analysis is a device integrating data collection, real-time monitoring, intelligent analysis, and remote control functions, mainly used to improve the operation and maintenance management efficiency of devices or systems. This terminal obtains device operating status, environmental parameters, and fault information through sensors or other data collection modules, and uses big data analysis, machine learning, or deep learning algorithms for intelligent diagnosis, predictive maintenance, and anomaly detection. Through edge computing or cloud computing, this terminal can evaluate the health status of devices in real time, automatically generate operation and maintenance suggestions, and actively give early warnings when potential risks are detected to avoid the occurrence of sudden failures. In addition, this terminal usually has remote operation and maintenance capabilities, and operation and maintenance personnel can monitor devices, adjust parameters, and schedule operation and maintenance through mobile terminals or cloud platforms, realizing refined management, reducing maintenance costs, and improving system reliability and operation and maintenance efficiency.

[0003] An optical fiber switch is one of the core devices in operation and maintenance intelligent terminals, mainly used in high-bandwidth, low-latency, and high-reliability network environments such as data centers, industrial operation and maintenance, smart grids, and intelligent transportation. Its function is to achieve high-speed data exchange and intelligent distribution among multiple devices through optical fiber communication technology, ensuring real-time monitoring and remote management of the operation and maintenance system. Compared with traditional Ethernet switches, optical fiber switches have higher transmission rates (such as 10G, 40G, 100G and above), lower signal attenuation, and stronger anti-interference capabilities, and are suitable for operation and maintenance scenarios that require large-scale data transmission and low-latency control. In addition, optical fiber switches support functions such as virtual local area network (VLAN), link aggregation control protocol (LACP), and Ethernet ring protection switching (ERPS), which can optimize the operation and maintenance network topology, improve communication stability, and provide redundant backups in critical operation and maintenance systems to ensure data reliability and service continuity.

[0004] The prior art has the following deficiencies:

[0005] The prior art usually maintains and manages fiber optic switches using a fixed operation and maintenance cycle, that is, performing routine inspections, troubleshooting, and maintenance work at preset time intervals (such as monthly, quarterly, or annually) to reduce the equipment failure rate and ensure the long-term stable operation of the fiber optic communication network. However, when there are potential hidden dangers in the fiber optic link but the fixed-cycle maintenance strategy is still used, serious consequences may occur. As a core network device, if the fiber optic switch fails to detect link failures in a timely manner, it may cause the paralysis of key communication links such as data centers, enterprise networks, industrial control systems, and smart grids, causing a serious impact on business continuity. In the telecommunications operator network, link anomalies may lead to the interruption of user networks in a large area, resulting in large-scale service failures and economic losses. Therefore, relying solely on a fixed operation and maintenance cycle is difficult to meet the high reliability requirements. It is urgent to introduce intelligent monitoring and predictive maintenance to achieve real-time fault warning and dynamic maintenance strategies, improve the stability and operation efficiency of fiber optic switches, avoid the lag problem of the fixed operation and maintenance cycle, ensure that network devices operate in an efficient and healthy state, and significantly reduce network interruptions or service failures caused by failure to detect problems in a timely manner, thereby improving the reliability, stability, and operation efficiency of the network to solve the problems in the above background technology.

[0006] The above information disclosed in the background art section is only used to enhance the understanding of the background of the present disclosure. Therefore, it may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0007] The object of the present invention is to provide an operation and maintenance intelligent terminal based on data analysis. By continuously collecting and processing the operation parameter data of fiber optic switches and combining the intelligent evaluation of feature engineering and machine learning models, the operation and maintenance intelligent terminal can accurately identify potential hidden dangers in the fiber optic link, dynamically adjust the maintenance cycle according to the severity of the link anomaly, respond to fault risks in a timely manner, effectively improve the fault warning ability and maintenance efficiency of fiber optic switches, avoid the lag problem of the fixed operation and maintenance cycle, ensure that network devices operate in an efficient and healthy state, significantly reduce network interruptions or service failures caused by failure to detect problems in a timely manner, thereby improving the reliability, stability, and operation efficiency of the network to solve the problems in the above background technology.

[0008] To achieve the above object, the present invention provides the following technical solution: An operation and maintenance intelligent terminal based on data analysis, including an initial operation and maintenance cycle setting module, a real-time data collection and monitoring module, a data preprocessing and feature extraction module, an intelligent evaluation and prediction module, and a dynamic maintenance strategy adjustment module;

[0009] The initial operation and maintenance cycle setting module sets an initial operation and maintenance cycle for the fiber optic switch, and performs routine maintenance, inspections, and optimizations according to the established cycle after the fiber optic switch runs to ensure the normal operation and stable performance of the fiber optic switch;

[0010] Real-time data acquisition and monitoring module, during the operation of the fiber optic switch, continuously acquires the operation parameter data of the fiber optic switch in real time, establishes a real-time data stream of the device status, ensures that all potential abnormal information can be captured in a timely manner, and provides raw data support for subsequent data processing and intelligent diagnosis;

[0011] Data preprocessing and feature extraction module, performs preprocessing operations on the acquired operation parameter data to ensure the accuracy and consistency of the data. Based on the preprocessed data, extracts the key features reflecting the abnormal health status of the fiber optic link. Under the monitoring window, further analyzes the extracted key features through feature engineering techniques to provide input for the machine learning model, ensuring that the data input into the machine learning model has high discrimination, so as to improve the accuracy of detecting and predicting potential abnormalities in the fiber optic link;

[0012] Intelligent evaluation and prediction module, sends the features processed by feature engineering into a machine learning model that has been pre-trained based on historical data, uses the machine learning model to comprehensively evaluate and predict the analyzed features, and automatically detects the operation status of the fiber optic link;

[0013] Dynamic maintenance strategy adjustment module, when detecting potential abnormalities in the operation status of the fiber optic link, intelligently and dynamically adjusts the periodic maintenance strategy of the fiber optic switch according to the severity of the potential abnormalities in the fiber optic link, dynamically shortens the actual operation and maintenance cycle, and timely detects link fault problems.

[0014] Preferably, the initial operation and maintenance cycle refers to the first maintenance time interval comprehensively set for the fiber optic switch based on maintenance records, fault statistics, and equipment life evaluation data, providing a basic maintenance plan for the operation and maintenance intelligent terminal. In the absence of intelligent monitoring or predictive maintenance, it ensures that the device will not have unpredictable failures due to long-term lack of maintenance.

[0015] Preferably, based on the preprocessed data, extracts the key features reflecting the abnormal health status of the fiber optic link. Among them, the extracted features include the rate of power attenuation of the optical signal during transmission and the non-linear growth of the bit error rate with the transmission distance. Under the monitoring window, further analyzes the rate of power attenuation of the optical signal during transmission and the non-linear growth of the bit error rate with the transmission distance through feature engineering techniques, respectively generates the optical power attenuation factor and the non-linear rise factor of the bit error rate, quantifies the abnormal operation status of the fiber optic link through the optical power attenuation factor and the non-linear rise factor of the bit error rate, and provides input for the machine learning model to improve the accuracy of detecting and predicting fiber optic link abnormalities.

[0016] Preferably, the optical power attenuation factor and the non - linear rise factor of the bit error rate after feature engineering processing are fed into a machine learning model pre - trained based on historical data. A link anomaly index is generated through the machine learning model, and the machine learning model is used to comprehensively evaluate and predict the analyzed link anomaly index for the automatic detection of the operating state of the optical fiber link.

[0017] Preferably, when evaluating and predicting the operating state of the optical fiber link through a trained machine learning model, the link anomaly index generated is compared and analyzed with a pre - set reference threshold of the link anomaly index for the automatic detection of the operating state of the optical fiber link. The specific detection steps are as follows:

[0018] If the link anomaly index is greater than the reference threshold of the link anomaly index, a link anomaly signal is generated, indicating that there are potential anomalies in the health state of the optical fiber link; if the link anomaly index is less than or equal to the reference threshold of the link anomaly index, a link normal signal is generated, indicating that the optical fiber link is operating efficiently and healthily.

[0019] Preferably, under the monitoring window, through feature engineering techniques, the rate of power attenuation of the optical signal during transmission is further analyzed. The specific steps for generating the optical power attenuation factor are as follows:

[0020] First, for any position L on the optical fiber link the measured optical power is non - linearly transformed to calculate the instantaneous attenuation rate. The calculation expression is: where: L represents the instantaneous optical power attenuation rate at position on the optical fiber link, L represents the real - time measured optical power at position on the optical fiber link, L represents the derivative of the optical fiber length which reflects the instantaneous change rate of the optical power with distance. The negative sign ensures that when the optical power attenuates, the instantaneous optical power attenuation rate

[0021] is positive; After obtaining the instantaneous optical power attenuation rate along the optical fiber link, an integrated attenuation index, that is, the optical power attenuation factor, is constructed by using integration and non - linear transformation. The constructed calculation formula is: where: and are the starting and ending positions of the optical fiber link respectively, is an exponential parameter used to amplify the contribution of high attenuation rate values, reflecting the non - linear effect of the intensifying anomaly, is the sensitivity coefficient, through the exponential function Further magnify the impact of the instantaneous decay rate when it suddenly increases locally to highlight abnormal phenomena.

[0022] Preferably, under the monitoring window, the specific steps for further analyzing the non - linear growth of the bit error rate with the transmission distance through feature engineering technology to generate the bit error rate non - linear increase factor are as follows:

[0023] Within the monitoring window, first, analyze the change of the bit error rate with the transmission distance x by local differential analysis to capture the instantaneous characteristics of the non - linear growth of the curve. The specific method is to calculate the first - order derivative and the second - order derivative of the bit error rate, and then define the local curvature function. The expression of the local curvature function is: , where: is the local curvature of the bit error rate of the optical fiber link with respect to the transmission distance x and is used to describe the degree of bending of the bit error rate curve at a certain point. is the first - order derivative of the bit error rate, representing the instantaneous change rate of the bit error rate with respect to the transmission distance. is the second - order derivative of the bit error rate, representing the acceleration of the instantaneous change rate and reflecting the "bending" trend of the increase in the bit error rate. The denominator is used for normalization to eliminate the influence of data scale on the curvature calculation.

[0024] After obtaining the local curvature function, further aggregate the local abnormal conditions within the entire monitoring window to generate the final bit error rate non - linear increase factor. The generation expression of the bit error rate non - linear increase factor is: , where: represents the bit error rate non - linear increase factor. represents the transmission distance interval within the monitoring window. and represent the start point and the end point of the transmission distance interval within the monitoring window respectively. is the absolute value of the local curvature to ensure that the contribution of each point to the abnormality is positive. is the power - exponent parameter used to enhance the sensitivity to local extreme abnormal values.

[0025] Preferably, when it is sensed that there are abnormal hidden dangers in the operation state of the optical fiber link, according to the severity of the abnormal hidden dangers of the optical fiber link, the periodic maintenance strategy of the optical fiber switch is intelligently and dynamically adjusted. The specific steps are as follows:

[0026] When it is sensed that there are abnormal hidden dangers in the optical fiber link, first calculate the exceeding amplitude of the link abnormal index relative to the link abnormal index reference threshold to quantify the severity of the optical fiber link abnormality. The calculation formula is as follows: , where: is the abnormal index of the optical fiber link after the evaluation of the machine learning model, reflecting the health degree of the current link operation status, is the reference threshold of the link abnormal index, that is, the standard value when the optical fiber link is in normal operation, is the abnormal amplitude of the link, used to measure the deviation degree of the link abnormal index relative to the reference threshold, and the index parameter is used to control the response sensitivity to the abnormal severity;

[0027] After quantifying the severity of the potential abnormal risks of the optical fiber link, the actual operation and maintenance cycle of the optical fiber switch is dynamically adjusted according to the abnormal amplitude of the link, ensuring that the maintenance frequency can respond to the changes in the link health status in a timely manner. The adjustment formula is as follows: , where: is the actually adjusted operation and maintenance cycle, used to determine the time of the next maintenance, is the initial operation and maintenance cycle, is the dynamic adjustment coefficient, used to control the adjustment amplitude.

[0028] In the above technical solution, the technical effects and advantages provided by the present invention are as follows:

[0029] By continuously collecting and processing the operation parameter data of the optical fiber switch, and combining with the intelligent evaluation of feature engineering and machine learning model, the operation and maintenance intelligent terminal can accurately identify the potential abnormal risks of the optical fiber link, dynamically adjust the maintenance cycle according to the severity of the link abnormality, respond to the fault risk in a timely manner, effectively improve the fault warning ability and maintenance efficiency of the optical fiber switch, avoid the lag problem of the fixed operation and maintenance cycle, ensure that the network device runs in an efficient and healthy state, and significantly reduce the network interruption or service failure caused by the failure to detect problems in time, thereby improving the reliability, stability and operation efficiency of the network. Brief Description of the Drawings

[0030] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required to be used in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained according to these drawings.

[0031] Figure 1 is the module schematic diagram of an operation and maintenance intelligent terminal based on data analysis according to the present invention. Detailed Embodiment

[0032] Example embodiments will now be described more fully with reference to the accompanying drawings. However, the example embodiments can be implemented in various forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the concept of the example embodiments to those skilled in the art.

[0033] The present invention provides an Figure 1 operation and maintenance intelligent terminal based on data analysis as shown, including an initial operation and maintenance cycle setting module, a real-time data collection and monitoring module, a data preprocessing and feature extraction module, an intelligent evaluation and prediction module, and a dynamic maintenance strategy adjustment module;

[0034] The initial operation and maintenance cycle setting module sets an initial operation and maintenance cycle (such as monthly, quarterly, or annually) for the fiber optic switch, and performs routine maintenance, inspection, and optimization according to the established cycle after the fiber optic switch runs, ensuring the normal operation and stable performance of the fiber optic switch;

[0035] The initial operation and maintenance cycle refers to the first maintenance time interval comprehensively set for the fiber optic switch based on maintenance records, fault statistics, and equipment life evaluation data. Its main function is to provide a basic maintenance plan to ensure that the equipment does not have unpredictable failures due to long-term lack of maintenance in the absence of intelligent monitoring or predictive maintenance.

[0036] This initial operation and maintenance cycle serves as the basis for maintenance management, ensuring that routine inspections can be carried out regularly even without abnormal warnings, providing a reliable benchmark maintenance framework, ensuring the regular maintenance of the equipment, and reducing the accumulation of potential faults that may be caused by long-term neglect.

[0037] Based on maintenance records, fault statistics, and equipment life evaluation data means that when setting the initial operation and maintenance cycle of the fiber optic switch, it is necessary to comprehensively consider the historical maintenance situation of the equipment, the probability of fault occurrence, and the expected service life of the equipment to ensure the rationality and scientificity of the maintenance cycle.

[0038] Maintenance records: Maintenance records refer to the information such as maintenance time, maintenance method, replaced parts, and optimized configuration recorded during the past operation and maintenance of the fiber optic switch and related network equipment. These data can reflect the maintenance frequency, maintenance effect, and whether there are repetitive faults of the equipment at different times. For example, if in the past maintenance records of a certain model of fiber optic switch, the optical interface of the optical module needs to be cleaned every 12 months, otherwise there will be a problem of optical power attenuation, then based on this data, the initial operation and maintenance cycle can be set to 12 months to avoid similar problems.

[0039] Fault Statistics: Fault statistics are based on large-scale equipment operation data to quantitatively analyze the failure rate, failure type, and failure severity of fiber optic switches under different environments, different loads, and different usage years. For example, in a data center network, after a certain brand of fiber optic switch has been continuously used for 12 months, the bit error rate (BER) has increased significantly, and more than 80% of the failures occur within 12 - 18 months. Then, a reasonable initial maintenance cycle should be set within 12 months to perform preventive maintenance before problems occur and reduce the risk of network interruption.

[0040] Equipment Life Evaluation Data: Equipment life evaluation is based on the physical life of the equipment, the law of performance degradation, the characteristics of material aging, and the lifecycle data recommended by the manufacturer to estimate the expected life of fiber optic switches and their core components (such as optical modules, power modules, fans, etc.) under different usage conditions. For example, the designed life of a certain model of optical module is 5 years, but in a high-temperature and high-load environment, the actual life may be shortened to 2 years. Then, the inspection and replacement frequency should be increased when it has been used for about 2 years to avoid unforeseen failures before the end of the equipment's life.

[0041] Through the comprehensive analysis of maintenance records, fault statistics, and equipment life evaluation data, the initial operation and maintenance cycle can be formulated more accurately, avoiding equipment anomalies caused by insufficient maintenance and preventing unnecessary operation and maintenance costs caused by over-frequent maintenance. In an intelligent operation and maintenance system, these historical data can also be further used for machine learning model training to achieve predictive maintenance, transforming the operation and maintenance cycle from a fixed mode to a dynamically adaptive optimization, improving the stability and service life of fiber optic switches.

[0042] The Real-time Data Acquisition and Monitoring Module continuously and real-time collects the operation parameter data of the fiber optic switch during its operation, establishing a real-time data stream of the equipment status to ensure that all potential abnormal information can be captured in a timely manner, providing raw data support for subsequent data processing and intelligent diagnosis;

[0043] Through embedded sensors, digital optical monitoring (DOM) of optical modules, SNMP protocol, and other network monitoring tools, continuously and real-time collect key parameter data. High-frequency and continuous data collection can help detect tiny change trends and then early warn of possible problems.

[0044] The Data Preprocessing and Feature Extraction Module performs preprocessing operations on the collected operation parameter data to ensure the accuracy and consistency of the data. Based on the preprocessed data, key features reflecting the abnormal health status of the fiber optic link are extracted. Under the monitoring window, the extracted key features are further analyzed through feature engineering techniques to provide input for the machine learning model, ensuring that the data input into the machine learning model has high discrimination to improve the accuracy of detecting and predicting potential anomalies in the fiber optic link;

[0045] Based on the preprocessed data, key features reflecting the abnormal health status of the optical fiber link are extracted. Among them, the extracted features include the rate of power attenuation of the optical signal during transmission and the non-linear growth of the bit error rate with the transmission distance. Under the monitoring window, through feature engineering techniques, the rate of power attenuation of the optical signal during transmission and the non-linear growth of the bit error rate with the transmission distance are further analyzed, and the optical power attenuation factor and the non-linear bit error rate increase factor are generated respectively. The abnormal operating state of the optical fiber link is quantified by the optical power attenuation factor and the non-linear bit error rate increase factor, and input is provided for the machine learning model to improve the accuracy of optical fiber link anomaly detection and prediction.

[0046] During the operation of the optical fiber switch, if the rate of power attenuation of the optical signal in the optical fiber link rises abnormally during transmission, this usually indicates potential abnormal hidden dangers in the health status of the optical fiber link. Under normal circumstances, the optical signal will have a certain attenuation due to its inherent physical properties during transmission, but this attenuation should be maintained within an expected stable range; once an abnormal increase occurs, it may indicate that the optical fiber link is affected by factors such as fiber aging, connector contamination, excessive bending, physical damage, or other environmental factors, resulting in increased signal loss. Such abnormal changes may further lead to an increase in the bit error rate, data transmission errors, and even cause the entire link to fail, and in severe cases, may cause communication interruption and network paralysis. Therefore, monitoring the abnormal changes in the optical signal power attenuation rate is crucial for timely warning and taking maintenance measures.

[0047] Under the monitoring window, the specific steps to generate the optical power attenuation factor by further analyzing the rate of power attenuation of the optical signal during transmission through feature engineering techniques are as follows:

[0048] First, perform a non-linear transformation on the optical power L measured at any position on the optical fiber link to calculate the instantaneous attenuation rate. The calculation expression is: where: represents the instantaneous optical power attenuation rate at position L on the optical fiber link. Specifically, it reflects the attenuation rate of the optical signal power per unit length (or distance). represents the real-time measured optical power at position L on the optical fiber link. is the natural logarithm of the optical power, aiming to convert the multiplicative attenuation effect into an additive form for easy processing. represents the derivative of the optical fiber length L , reflecting the instantaneous change rate of the optical power with distance. The negative sign ensures that when the optical power attenuates, the instantaneous optical power attenuation rate is positive;

[0049] Convert the multiplicative attenuation effect into an additive form, with the function of:

[0050] By taking the natural logarithm of the optical signal power value, the multiplicative effect of the optical power attenuation with distance in the optical fiber link is converted into an additive effect. Physically, the attenuation of an optical signal is usually multiplicative, that is, the signal power decays by a fixed ratio as the optical fiber transmission distance increases. This attenuation shows an exponential decrease. By taking the natural logarithm, this exponential attenuation is converted into an additive form, making it a linear change. The purpose of this is to make the signal attenuation more convenient for analysis and processing, because the additive effect can be studied through simple difference or derivative operations, while the multiplicative effect usually increases the computational complexity. When calculating the attenuation rate, through the conversion of the natural logarithm, the rate and pattern of attenuation can be studied more directly.

[0051] The function of this step is to convert the original optical power data into a local attenuation rate curve, capture the instantaneous attenuation behavior of the optical signal during transmission, lay a foundation for subsequent comprehensive calculations, and be able to sensitively reflect local abnormal attenuation phenomena.

[0052] After obtaining the instantaneous optical power attenuation rate along the optical fiber link then use integration and non-linear transformation to construct a comprehensive attenuation index, that is, the optical power attenuation factor. The constructed calculation formula is: , where: is the optical power attenuation factor, and are the starting and ending positions of the optical fiber link respectively, is an exponential parameter used to amplify the contribution of high attenuation rate values and reflect the non-linear effect of abnormal aggravation, is the sensitivity coefficient, and through the exponential function further amplifies the influence of the instantaneous attenuation rate when it suddenly increases locally, highlighting abnormal phenomena;

[0053] The function of this step is to cumulatively evaluate the local attenuation rate within the entire optical fiber link, use integral operation to converge abnormal signals everywhere, and then through non-linear amplification, make local abnormalities fully reflected in the comprehensive index to achieve precise quantification of the abnormal health state of the optical fiber link.

[0054] From the optical power attenuation factor, it can be seen that under the monitoring window, the larger the performance value of the optical power attenuation factor generated by further analyzing the rate of optical signal power attenuation during transmission through feature engineering technology, the higher the risk of abnormal hidden dangers in the health status of the optical fiber link. By comprehensively evaluating the rate and change trend of power attenuation in the optical fiber link, the optical power attenuation factor can reflect the change in signal quality during the transmission process of the optical fiber link. When the value of the optical power attenuation factor is large, it usually means that significant power attenuation or accelerated attenuation has occurred in certain areas of the link, which may be caused by fiber aging, damage, poor connection or other external factors, indicating that there is a potential fault risk in the link. On the contrary, when the value of the optical power attenuation factor is small, it indicates that the optical signal transmission of the link is stable and the attenuation rate is relatively gentle, indicating that the health status of the optical fiber link is good and the abnormal hidden dangers are small. Therefore, as a key indicator, the optical power attenuation factor can help operation and maintenance personnel identify the health problems of the optical fiber link in a timely manner.

[0055] During the operation of the optical fiber switch, if the bit error rate of the optical fiber link shows a non-linear upward trend with the increase of the transmission distance, it indicates that there may be abnormal hidden dangers in the link. Usually, the bit error rate will increase to a certain extent with the increase of the distance, but this increase should be stable and gradual; once there is a non-linear and sharp increase in the bit error rate, it may mean that there are problems such as physical bending of the optical fiber, connector contamination, optical module aging or abnormal optical fiber attenuation, which will cause the signal quality to deteriorate rapidly, thus triggering transmission errors. Timely capturing of this abnormal phenomenon is crucial for preventing potential faults, conducting targeted maintenance and ensuring the stability of the overall network communication.

[0056] Under the monitoring window, the specific steps for generating the bit error rate non-linear upward factor by further analyzing the non-linear growth of the bit error rate with the transmission distance through feature engineering technology are as follows:

[0057] Within the monitoring window, first, analyze the change of the bit error rate with the transmission distance x by local differential analysis to capture the instantaneous characteristics of the non-linear growth of the curve. The specific method is to calculate the first derivative and the second derivative of the bit error rate, and then define the local curvature function. The expression of the local curvature function is: , where: is the local curvature of the bit error rate of the optical fiber link with the change of the transmission distance x , which is used to describe the bending degree of the bit error rate curve at a certain point, is the first derivative of the bit error rate, representing the instantaneous change rate of the bit error rate with respect to the transmission distance, is the second derivative of the bit error rate, representing the acceleration of the instantaneous change rate, reflecting the "bending" trend of the increase of the bit error rate, and is used for normalization to eliminate the influence of data scale on the curvature calculation;

[0058] The function of this step is to convert the non-linear growth characteristic of the bit error rate into a curvature value in mathematics. A higher local curvature indicates a sharp increase in the local bit error rate, thus revealing potential abnormal hidden dangers in the optical fiber link.

[0059] After obtaining the local curvature function, further aggregate the local abnormal conditions within the entire monitoring window to generate the final non-linear increase factor of the bit error rate. The generation expression of the non-linear increase factor of the bit error rate is: , where: represents the non-linear increase factor of the bit error rate, represents the transmission distance interval within the monitoring window, and represent the start point and end point of the transmission distance interval within the monitoring window respectively, is the absolute value of the local curvature, ensuring that the contribution of each point to the abnormality is positive, is the power exponent parameter, which is used to enhance the sensitivity to local extreme outliers (for example, when taking , the parts with higher local curvature contribute more to the integral).

[0060] It can be seen from the non-linear increase factor of the bit error rate that under the monitoring window, the larger the value of the non-linear increase factor of the bit error rate generated by further analyzing the non-linear growth of the bit error rate with the transmission distance through feature engineering technology, the greater the possibility of abnormal hidden dangers existing in the health state of the optical fiber link. This is because the non-linear increase exponent of the bit error rate with the transmission distance captures the trend of the sharp increase in the bit error rate by quantifying the curvature change of the optical fiber link within the monitoring window. A higher non-linear increase factor of the bit error rate means that the change rate and acceleration of the bit error rate within a certain distance increase significantly. Usually, due to damage to the optical fiber link, degradation of the optical module, or other external factors, the signal quality drops sharply, which are all early signs of potential failures. On the contrary, when the non-linear increase factor of this bit error rate is low, it indicates that the growth of the bit error rate with distance is relatively gentle, the link health state is relatively stable, and there are no obvious abnormal hidden dangers. Therefore, the non-linear increase factor of the bit error rate can effectively reflect the health state of the optical fiber link, and as an important warning signal in the intelligent operation and maintenance system, it helps to detect and repair possible link problems in a timely manner.

[0061] The intelligent evaluation and prediction module sends the features processed by feature engineering into a machine learning model that has been trained in advance based on historical data, and uses the machine learning model to comprehensively evaluate and predict the analyzed features for automatic detection of the operating state of the optical fiber link;

[0062] Send the optical power attenuation factor and the non - linear rise factor of the bit error rate after feature engineering into a machine learning model that has been pre - trained based on historical data. Generate a link anomaly index through the machine learning model, and use the machine learning model to comprehensively evaluate and predict the analyzed link anomaly index for the automatic detection of the operating state of the optical fiber link.

[0063] The machine learning model pre - trained based on historical data refers to a prediction model constructed using supervised or semi - supervised learning algorithms after data pre - processing and feature engineering, using a large amount of past accumulated operation and maintenance records, fault statistics, equipment life data, and key feature data such as real - time collected optical power attenuation and bit error rate before officially putting it into real - time optical fiber link monitoring and anomaly detection. During the construction of this model, by inputting various parameters in historical data (such as the optical power attenuation factor and the non - linear rise factor of the bit error rate extracted through feature engineering), it learns and captures the characteristic patterns of the optical fiber link in normal operation and abnormal states, as well as the complex non - linear relationships between these characteristics and the fault risk. The model may adopt algorithms such as deep neural networks, random forests, and gradient - boosting decision trees, and continuously adjust internal parameters during repeated training and verification to achieve an accurate association between input features and the link anomaly index with the optimal mapping relationship. During the model training process, cross - validation, loss function optimization, and regularization techniques are also used to ensure that the model has high generalization ability and robustness when facing new data. After the trained model parameters are fixed, it can be used as a pre - trained "expert system" to quickly map the new data after feature engineering to a comprehensive evaluation index, the link anomaly index, during real - time monitoring, so as to automatically detect and predict the health status of the optical fiber link. The advantage of this pre - trained model is that it can identify subtle anomalies and potential fault signs that may be overlooked by traditional statistical methods through the complex laws contained in historical data, providing data support and decision - making basis for subsequent predictive maintenance.

[0064] In practical applications, the significance of this pre - trained machine learning model is particularly important because it endows the optical fiber link monitoring system with the characteristics of adaptability and intelligence. Specifically, after the model is trained on historical data, it is deployed on the monitoring platform. When real - time data streams in, the system first pre - processes the original sensor data, then extracts the quantified key features through feature engineering, such as the optical power attenuation factor and the non - linear rise factor of the bit error rate, and then inputs these features into the pre - trained model for comprehensive evaluation. Based on the historical patterns and anomaly features learned internally, the model quickly outputs a link anomaly index, which not only reflects the degree of abnormality of the current data state but also can predict the possible future fault risks.

[0065] The machine learning model is not specifically limited herein, as long as it can implement comprehensive analysis of the optical power attenuation factor and the non-linear rise factor of the bit error rate to generate a link anomaly index is acceptable. To implement the technical solution of the present invention, the present invention provides a specific implementation method; the link anomaly index is generated according to the following calculation formula: , where , are respectively the preset proportionality coefficients of the optical power attenuation factor and the non-linear rise factor of the bit error rate , and , are both greater than 0. The preset proportionality coefficients refer to the constant coefficients and set to effectively combine and adjust different input variables (such as the optical power expression index and the non-linear rise factor of the bit error rate ) when calculating the link anomaly index. The role of these preset proportionality coefficients is to weight the importance of different factors according to the actual situation, so that their contribution to the final result matches the actually observed impact. For example, is the preset proportionality coefficient related to the optical power expression index , and is the coefficient related to the non-linear rise factor of the bit error rate . By adjusting these preset proportionality coefficients, the relative influence of the two in the calculation of the link anomaly index can be changed to ensure that the model can accurately reflect the real situation of the health status of the optical fiber link. Especially in actual operation and maintenance, these preset proportionality coefficients may be dynamically adjusted according to the characteristics of different devices, environmental factors, fault types, etc. to achieve more accurate anomaly detection and prediction.

[0066] It can be seen from the link anomaly index that under the monitoring window, the larger the value of the optical power attenuation factor generated by further analyzing the rate of optical power attenuation during the transmission of optical signals through feature engineering technology, and the larger the value of the non-linear rise factor of the bit error rate generated by further analyzing the non-linear growth of the bit error rate with the transmission distance through feature engineering technology, that is, the larger the value of the link anomaly index generated when evaluating and predicting the operating status of the optical fiber link through a trained machine learning model, the higher the risk of abnormal hidden dangers in the health status of the optical fiber link, and vice versa, the lower the risk of abnormal hidden dangers in the health status of the optical fiber link.

[0067] When evaluating and predicting the operating status of an optical fiber link through a trained machine learning model, the link anomaly index generated is compared with a pre-set reference threshold of the link anomaly index for automated detection of the operating status of the optical fiber link. The specific detection steps are as follows:

[0068] If the link anomaly index is greater than the reference threshold of the link anomaly index, a link anomaly signal is generated, indicating that there are potential hidden dangers in the health status of the optical fiber link; if the link anomaly index is less than or equal to the reference threshold of the link anomaly index, a link normal signal is generated, indicating that the optical fiber link is operating efficiently and healthily.

[0069] The dynamic maintenance strategy adjustment module, when perceiving potential hidden dangers in the operating status of the optical fiber link, intelligently and dynamically adjusts the periodic maintenance strategy of the optical fiber switch according to the severity of the potential hidden dangers of the optical fiber link, dynamically shortens the actual operation and maintenance cycle, and timely detects link failure problems;

[0070] When perceiving potential hidden dangers in the operating status of the optical fiber link, intelligently and dynamically adjusts the periodic maintenance strategy of the optical fiber switch according to the severity of the potential hidden dangers of the optical fiber link. The specific steps are as follows:

[0071] When perceiving potential hidden dangers in the optical fiber link, first calculate the excess amplitude of the link anomaly index relative to the reference threshold of the link anomaly index to quantify the severity of the optical fiber link anomaly. The calculation formula is as follows: , where: is the optical fiber link anomaly index evaluated by the machine learning model, reflecting the health degree of the current link operating status, is the reference threshold of the link anomaly index, that is, the standard value when the optical fiber link is in normal operation, is the link anomaly excess amplitude, used to measure the deviation degree of the link anomaly index relative to the reference threshold, and the exponential parameter is used to control the response sensitivity to the severity of the anomaly. When the anomaly index is slightly exceeded, the impact is small, while when the anomaly exceeds a large margin, the corresponding adjustment margin will be amplified;

[0072] The function of this step is to accurately quantify the severity of the abnormal situation, ensure that the system can dynamically adapt to different levels of abnormal situations, avoid over-frequent adjustment of the operation and maintenance strategy, and at the same time ensure that it can respond quickly when the anomaly reaches a dangerous level.

[0073] After quantifying the severity of the potential hidden dangers of the optical fiber link, dynamically adjust the actual operation and maintenance cycle of the optical fiber switch according to the link anomaly excess amplitude to ensure that the maintenance frequency can respond in a timely manner to changes in the link health status. The adjustment formula is as follows: , where: It is the actually adjusted operation and maintenance cycle, which is used to determine the time of the next maintenance. It is the initial operation and maintenance cycle. It is the dynamic adjustment coefficient, which is used to control the adjustment range, ensuring that the adjustment range is small when the anomaly is small, and when the anomaly exceeds the standard significantly, the operation and maintenance cycle can be significantly shortened.

[0074] The function of this step is to make the operation and maintenance strategy have dynamic adaptability, ensuring that when the optical fiber link is in good health, the operation and maintenance cycle remains long, reducing unnecessary maintenance, and when the link anomaly intensifies, the system can actively shorten the maintenance cycle, intervene in advance, and prevent potential failures from evolving into serious problems.

[0075] Through the above operation and maintenance intelligent terminal solution based on data analysis, it is possible to achieve intelligent real-time monitoring and dynamic operation and maintenance control of the operation status of the optical fiber switch. By continuously collecting and processing the operation parameter data of the optical fiber switch, combined with the intelligent evaluation of feature engineering and machine learning models, the system can accurately identify potential abnormal hidden dangers of the optical fiber link, and dynamically adjust the maintenance cycle according to the severity of the link anomaly, and respond to the fault risk in a timely manner. This solution effectively improves the fault warning ability and maintenance efficiency of the optical fiber switch, avoids the lag problem of the fixed operation and maintenance cycle, ensures that the network device operates in an efficient and healthy state, and significantly reduces network interruptions or service failures caused by failure to detect problems in a timely manner, thereby improving the reliability, stability and operation efficiency of the network.

[0076] The above formulas are all dimensionless and take their numerical values for calculation. The formula is a formula obtained by collecting a large amount of data for software simulation to approximate the real situation. The preset parameters in the formula are set by those skilled in the art according to the actual situation.

[0077] The above is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of changes or substitutions, which should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claimed rights.

[0078] Only some exemplary embodiments of the present invention have been described above by way of illustration. Undoubtedly, for those of ordinary skill in the art, the described embodiments can be modified in various different ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the protection scope of the claims of the present invention.

Claims

1. An operation and maintenance intelligent terminal based on data analysis, characterized in that: It includes the initial operation and maintenance cycle setting module, real-time data collection and monitoring module, data preprocessing and feature extraction module, intelligent evaluation and prediction module and dynamic maintenance strategy adjustment module; The initial operation and maintenance cycle setting module sets an initial operation and maintenance cycle for the fiber optic switch, and performs routine maintenance, inspection and optimization according to the established cycle after the fiber optic switch is put into operation; Real-time data collection and monitoring module, during the operation of the fiber optic switch, continuously collects the operating parameter data of the fiber optic switch in real time and establishes a real-time data stream of the device status; The data preprocessing and feature extraction module performs preprocessing operations on the collected operating parameter data to ensure the accuracy and consistency of the data. Based on the preprocessed data, the module extracts the key features that reflect the abnormal health status of the optical fiber link. Under the monitoring window, the module further analyzes the extracted key features through feature engineering technology to provide input for the machine learning model. The intelligent evaluation and prediction module feeds the features processed by feature engineering into a machine learning model that has been trained based on historical data in advance, and uses the machine learning model to comprehensively evaluate and predict the analyzed features, and automatically detect the operating status of the optical fiber link; The dynamic maintenance strategy adjustment module, when sensing the abnormal hidden dangers in the operation status of the optical fiber link, will intelligently and dynamically adjust the periodic maintenance strategy of the optical fiber switch according to the severity of the abnormal hidden dangers of the optical fiber link; Under the monitoring window, after further analysis of the extracted key features through feature engineering technology, the optical power attenuation factor and bit error rate nonlinear increase factor processed by feature engineering are sent to the machine learning model trained in advance based on historical data. The link anomaly index is generated through the machine learning model, and the machine learning model is used to comprehensively evaluate and predict the analyzed link anomaly index to automatically detect the operating status of the optical fiber link.

2. According to claim 1, the intelligent operation and maintenance terminal based on data analysis is characterized in that: The initial operation and maintenance cycle refers to the first maintenance time interval comprehensively set for the fiber optic switch based on maintenance records, failure statistics and equipment life assessment data.

3. The intelligent operation and maintenance terminal based on data analysis according to claim 1, characterized in that: Based on the preprocessed data, key features reflecting abnormal health status of optical fiber links are extracted. The extracted key features include the rate of power attenuation of optical signals during transmission and the nonlinear growth of bit error rate with transmission distance. Under the monitoring window, the extracted key features are further analyzed by feature engineering technology, specifically: Through feature engineering technology, the rate of power attenuation of optical signals during transmission and the nonlinear growth of bit error rate with transmission distance are further analyzed, and the optical power attenuation factor and the bit error rate nonlinear increase factor are generated respectively. The optical power attenuation factor and the bit error rate nonlinear increase factor are used to quantify the abnormal operating status of the optical fiber link, and provide input for the machine learning model to improve the accuracy of optical fiber link anomaly detection and prediction.

4. The operation and maintenance intelligent terminal based on data analysis according to claim 1, characterized in that: The link anomaly index generated by evaluating and predicting the operating status of the optical fiber link through the trained machine learning model is compared and analyzed with the pre-set link anomaly index reference threshold to automatically detect the operating status of the optical fiber link. The specific detection steps are as follows: If the link abnormality index is greater than the link abnormality index reference threshold, a link abnormality signal is generated, indicating that there are abnormal risks in the health status of the optical fiber link; if the link abnormality index is less than or equal to the link abnormality index reference threshold, a link normal signal is generated, indicating that the optical fiber link is operating efficiently and healthily.

5. The intelligent operation and maintenance terminal based on data analysis according to claim 3 is characterized in that: In the monitoring window, the rate of power attenuation of the optical signal during transmission is further analyzed through feature engineering technology. The specific steps for generating the optical power attenuation factor are as follows: First, any location on the fiber link L The optical power measured at A nonlinear transformation is performed to calculate the instantaneous decay rate. The expression for the calculation is: ,in: Indicates the position on the optical fiber link L The instantaneous optical power attenuation rate at Indicates the location of the optical fiber link L Real-time measurement of optical power at Indicates the length of the optical fiber L The derivative of reflects the instantaneous rate of change of optical power with distance. The negative sign ensures that when the optical power decays, the instantaneous optical power decay rate is a positive value; In order to obtain the instantaneous optical power attenuation rate along the optical fiber link Finally, the comprehensive attenuation index, namely the optical power attenuation factor, is constructed by using integration and nonlinear transformation. The calculation formula is: ,in: is the optical power attenuation factor, and are the starting and ending positions of the optical fiber link, is an exponential parameter used to amplify the contribution of high decay rate values, reflecting the nonlinear effect of abnormal aggravation. is the sensitivity coefficient, through the exponential function The impact of the local sudden increase in instantaneous decay rate is further amplified to highlight the abnormal phenomenon.

6. The intelligent operation and maintenance terminal based on data analysis according to claim 3, characterized in that: In the monitoring window, the nonlinear growth of the bit error rate with the transmission distance is further analyzed by feature engineering technology to generate the nonlinear increase factor of the bit error rate. The specific steps are as follows: In the monitoring window, we first analyze the bit error rate as a function of transmission distance. x The local differential analysis is performed on the changes of to capture the instantaneous characteristics of the nonlinear growth of the curve. The specific method is to calculate the first-order derivative and second-order derivative of the bit error rate, and then define the local curvature function. The expression of the local curvature function is: ,in: The bit error rate of the optical fiber link varies with the transmission distance. x The local curvature of the change is used to describe the curvature of the bit error rate curve at a certain point. is the first-order derivative of the bit error rate, which indicates the instantaneous rate of change of the bit error rate relative to the transmission distance. is the second-order derivative of the bit error rate, indicating the acceleration of the instantaneous rate of change, reflecting the "bending" trend of the bit error rate increase. The denominator Used for normalization to eliminate the impact of data scale on curvature calculation; After obtaining the local curvature function, the local abnormal conditions in the entire monitoring window are further aggregated to generate the final bit error rate nonlinear increase factor. The generation expression of the bit error rate nonlinear increase factor is: ,in: It represents the nonlinear increase factor of bit error rate, Indicates the transmission distance interval within the monitoring window. and They represent the starting point and end point of the transmission distance interval within the monitoring window respectively. is the absolute value of the local curvature, ensuring that each point contributes positively to the anomaly, is the power exponent parameter, which is used to enhance the sensitivity to local extreme outliers.

7. The intelligent operation and maintenance terminal based on data analysis according to claim 4, characterized in that: When an abnormal hidden danger is detected in the operation status of the optical fiber link, the periodic maintenance strategy of the optical fiber switch is intelligently and dynamically adjusted according to the severity of the abnormal hidden danger of the optical fiber link. The specific steps are as follows: When an abnormality risk is detected in a fiber link, the excess of the link abnormality index relative to the link abnormality index reference threshold is first calculated to quantify the severity of the fiber link abnormality. The calculation formula is as follows: ,in: It is the fiber link anomaly index evaluated by the machine learning model, reflecting the health of the current link operation status. It is the reference threshold of the link abnormality index, that is, the standard value of the fiber link in normal operation. is the link anomaly excess amplitude, which is used to measure the deviation of the link anomaly index from the reference threshold. The index parameter Used to control the sensitivity of the response to the severity of the anomaly; After quantifying the severity of the fiber link abnormality risk, the actual operation and maintenance cycle of the fiber switch is dynamically adjusted according to the link abnormality excess range to ensure that the maintenance frequency can respond to changes in the link health status in a timely manner. The adjustment formula is as follows: ,in: It is the actual adjusted operation and maintenance cycle, which is used to determine the time of the next maintenance. It is the initial operation and maintenance cycle. It is a dynamic adjustment coefficient, which is used to control the adjustment range.

Images