Intelligent early warning method and device for optical power fluctuation based on wide fixed network FTTH (Fiber To The Home)

Abstract

The invention relates to the technical field of intelligent network detection, in particular to an intelligent early warning method and device for optical power fluctuation based on wide fixed network FTTH, and the method comprises the following steps: S1, data collection: collecting an optical power value, a device ID and a timestamp of an optical port device; s2, data preprocessing, including data cleaning, format conversion, data standardization and data enhancement; s3, data analysis, including data mining, statistical analysis, anomaly detection and trend analysis; s4, fault prediction, including feature extraction, model training, model verification and model optimization; s5, intelligent early warning, including real-time data acquisition, real-time prediction, multi-level early warning mechanism, automatic order sending and closed-loop feedback; and S6, data visualization and report generation, including real-time interface monitoring, historical data query and automatic report generation, solving the problems of untimely monitoring and low processing efficiency of wide fixed network-based FTTH optical power fluctuation in the prior art.

Description

Technical Field

[0001] The present invention relates to the technical field of network intelligent detection, and in particular to an intelligent early warning method and device based on broadband fixed-line FTTH optical power fluctuation. Background Art

[0002] FTTH, short for "Fiber to the Home", is a technology that connects Internet services directly to users' homes through optical fibers. Compared with traditional copper-wire broadband, FTTH can provide faster transmission speeds and more stable network connections. Broadband fixed networks are part of access technologies such as FTTH (Fiber to the Home), and are mainly used to provide users with data, voice, video and other services. In FTTH networks, when optical fibers transmit signals, the optical power will fluctuate with factors such as the transmission distance of the signal, network load, and device status. The fluctuation of optical power will affect the stability of the network and the quality of the signal. If it cannot be handled in a timely manner, it may cause service interruption for a large number of users, increase operation and maintenance costs, and affect user experience.

[0003] At present, operators mainly manage the problem of optical power fluctuation through regular inspections and manual monitoring, and combine simple network management systems (NMS) to monitor optical power. When optical power abnormalities are detected, the system will generate a warning message, and the operation and maintenance personnel need to analyze and handle the faults according to the warning information. However, since the above methods rely on manual monitoring and regular inspections, it is difficult to detect the problem of optical power fluctuation in time, and since fault analysis and processing also rely on manual work, the response speed to faults is slow and inefficient. In summary, the existing technology has the problems of untimely monitoring and low processing efficiency for optical power fluctuations based on broadband fixed-line FTTH. Summary of the invention

[0004] In order to solve the problems existing in the prior art, the present application provides a method and device for intelligent early warning of optical power fluctuation based on broadband fixed-line FTTH network.

[0005] In a first aspect, the present invention provides an intelligent early warning method for optical power fluctuation based on broadband fixed-line FTTH network, which adopts the following technical solution:

[0006] An intelligent early warning method for optical power fluctuation based on broadband fixed-line FTTH network includes the following steps:

[0007] S1: Data collection, including collecting the optical power value, device ID and timestamp of the optical port device;

[0008] S2: Data preprocessing, including data cleaning, format conversion, data standardization and data enhancement;

[0009] S3: Data analysis, including data mining, statistical analysis, anomaly detection, and trend analysis;

[0010] S4: Fault prediction, including feature extraction, model training, model validation, and model optimization;

[0011] S5: Intelligent early warning, including real-time data collection, real-time prediction, multi-level early warning mechanism, automatic dispatching and closed-loop feedback;

[0012] S6: Data visualization and report generation, including real-time monitoring interface, historical data query and automatic report generation.

[0013] Preferably, in S1, the optical power value, device ID and timestamp are collected from each optical port device in the network through the SNMP protocol, including PON optical power collection and optical modem optical power collection; the user's offline records are collected from the business system, the data related to the optical power fluctuation are analyzed therefrom, and the user's device usage habits are integrated to form a complete data view.

[0014] Preferably, in S2, data cleaning includes: cleaning the collected data, including deleting duplicate data, correcting erroneous data, filling missing data, ensuring the integrity and accuracy of the data, and identifying and correcting abnormal values ​​through abnormal data algorithms, such as Z-score, IQR and Boxpiot; format conversion includes: converting data from different sources into a unified format and using standardized format data; data standardization includes: standardizing the optical power value, using normalization or Z-score standardization method, and unifying the measurement unit of the data; data enhancement includes: enhancing the data through interpolation and smoothing methods.

[0015] Preferably, in S3, data mining includes: using data mining algorithms, such as classification algorithms, clustering algorithms, and regression algorithms, to extract the changing trends and abnormal fluctuations of optical power values ​​from a large amount of data, and identifying different types of optical power fluctuation patterns through clustering analysis; statistical analysis includes: performing statistical analysis on optical power values, calculating mean, variance, and standard deviation indicators, determining the range of normal optical power values, and identifying the normal fluctuation range and abnormal fluctuations of optical power through probability distribution analysis; anomaly detection includes: using anomaly detection algorithms such as isolation forests and support vector machines to identify optical power values ​​that exceed the normal range and determine potential faulty optical ports; trend analysis includes: performing time series analysis on optical power data, identifying long-term trends and periodic fluctuations, and predicting future optical power changing trends.

[0016] Preferably, in S4, feature extraction includes: extracting key features from the analysis data, such as the optical power fluctuation amplitude, fluctuation frequency, fluctuation period, and fluctuation mutation point, and constructing a multidimensional feature space; model training includes: using machine learning algorithms, such as support vector machines, decision trees, and neural networks, to train historical data, establish a fault prediction model, and improve the training effect of the model through feature engineering; model verification includes: verifying the model through cross-validation and leave-one-out method, evaluating its prediction accuracy and robustness, and using confusion matrix and ROC curve evaluation indicators to quantify model performance; model optimization includes: optimizing the fault prediction model, using ensemble learning and hyperparameter tuning techniques to improve the generalization ability and prediction accuracy of the model.

[0017] Preferably, in S5, real-time data collection includes: continuously collecting new data in the network, using distributed data storage and processing technology for preprocessing and storage; real-time prediction includes: inputting new data into the fault prediction model, performing real-time prediction, outputting warning results, and real-time predicting the system's response to sudden changes in optical power to provide fault warnings; the multi-level warning mechanism includes: setting a multi-level warning mechanism according to the severity of the warning results, taking different treatment measures from general warnings to emergency warnings, and using a multi-level warning mechanism to distinguish different degrees of optical power fluctuations; automatic dispatching includes: automatically generating work orders through an artificial intelligence system based on the warning results, and dispatching them to relevant maintenance personnel. The work order system records the detailed process and results of each fault handling to provide data support for subsequent analysis; closed-loop feedback includes: after the maintenance personnel have handled the fault, they will feed back the processing results to the system. The system optimizes the fault prediction model based on the feedback results to achieve closed-loop management of fault handling.

[0018] Preferably, in S6, the real-time monitoring interface includes: providing a monitoring interface to display optical power data, warning information and processing status in real time; historical data query includes: supporting historical data query and analysis, and optimizing fault handling strategies; automatic report generation includes: the system automatically generates an optical power fluctuation analysis report, including data statistics, anomaly detection, fault prediction and processing results.

[0019] In a first aspect, the present invention provides a device for intelligent early warning of optical power fluctuation based on broadband fixed-line FTTH network, which adopts the following technical solution:

[0020] A device based on intelligent early warning of optical power fluctuation of broadband fixed-line FTTH network comprises a data acquisition module, a data preprocessing module, a data analysis module, a fault prediction module, an intelligent early warning module and a data visualization and report generation module which are connected in sequence.

[0021] The beneficial effects of the present invention are:

[0022] The present invention realizes intelligent monitoring of optical power fluctuations. Through big data analysis and artificial intelligence technology, it monitors and analyzes optical power fluctuations in real time, promptly discovers potential faults, automates fault warnings and dispatches, automatically generates fault warnings and dispatches them to maintenance personnel, improves fault handling efficiency, reduces operation and maintenance costs, and solves the problems of untimely monitoring and low processing efficiency for optical power fluctuations based on broadband fixed-line FTTH in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 This is a flow chart of an intelligent early warning method for optical power fluctuation based on broadband fixed-line FTTH in an embodiment of the present application;

[0024] Figure 2 It is a module diagram of an intelligent early warning device for optical power fluctuation based on broadband fixed-line FTTH in an embodiment of the present application. DETAILED DESCRIPTION

[0025] The following will be combined Figure 1-Figure 2 The present invention is further illustrated by the following embodiments.

[0026] This embodiment discloses an intelligent early warning method for optical power fluctuation based on broadband fixed-line FTTH network.

[0027] Reference Figure 1 The intelligent early warning method for optical power fluctuation based on broadband fixed-line FTTH includes the following steps:

[0028] S1: Data collection: Through the SNMP protocol, optical power values, device IDs, and timestamps are collected from each optical port device in the network, including PON optical power collection and optical modem optical power collection. These data provide a basis for subsequent big data analysis and fault prediction.

[0029] In addition to collecting the above-mentioned optical power data, it is also possible to collect multi-dimensional data such as the operating status of optical port equipment, environmental parameters such as humidity, temperature, network quality information, signal strength, optical signal-to-noise ratio, optical loss, wavelength, fiber link status information, optical power dynamic range, reflection loss, link quality, etc., to improve the accuracy and comprehensiveness of the analysis. In short, the amount of data that needs to be collected is determined by the scale and complexity of the network and the service quality requirements.

[0030] In addition, users’ offline records are collected from the business system to analyze data that may be related to optical power fluctuations. User device usage habits are integrated to form a complete data view.

[0031] In addition, it is necessary to set a reasonable acquisition frequency, such as once per second (1Hz), to ensure real-time performance. High-frequency acquisition can more accurately capture subtle fluctuations and sudden changes in optical power.

[0032] S2: Data preprocessing, including data cleaning, format conversion, data standardization and data enhancement.

[0033] Data cleaning: cleaning the collected data, including deleting duplicate data, correcting erroneous data, filling in missing data, ensuring the integrity and accuracy of the data, and identifying and correcting outliers through abnormal data algorithms, such as Z-score (standard deviation method), IQR (interquartile range method) and Boxp i ot (box plot).

[0034] Format conversion converts data from different sources into a unified format to facilitate subsequent processing and analysis, and uses standardized format data to improve the compatibility and scalability of the system.

[0035] Data standardization: standardize the optical power values ​​to eliminate the differences between devices with different optical ports, ensure the fairness and accuracy of data analysis, and use normalization or Z-score standardization methods to unify the measurement units of data.

[0036] Data enhancement, through interpolation, smoothing and other methods to enhance the data, reduce the impact of data missing and noise on the analysis results, the enhanced data can better reflect the actual fluctuation of optical power.

[0037] S3: Data analysis, including data mining, statistical analysis, anomaly detection, and trend analysis.

[0038] Data mining uses data mining algorithms, such as classification algorithms, clustering algorithms, and regression algorithms, to extract valuable information from large amounts of data, such as the changing trends and abnormal fluctuations of optical power values. In particular, cluster analysis can be used to identify different types of optical power fluctuation patterns.

[0039] Statistical analysis: Statistical analysis is performed on the optical power values ​​to calculate the mean, variance, standard deviation and other indicators to determine the range of normal optical power values. Through probability distribution analysis, the normal fluctuation range and abnormal fluctuation of optical power can be identified.

[0040] Anomaly detection uses anomaly detection algorithms to identify optical power values ​​that exceed the normal range and determine potential faulty optical ports. Algorithms such as isolation forest and support vector machines are used to improve the accuracy and reliability of anomaly detection.

[0041] Trend analysis: perform time series analysis on optical power data to identify long-term trends and periodic fluctuations and predict future optical power trends.

[0042] S4: Fault prediction, including feature extraction, model training, model validation, and model optimization.

[0043] Feature extraction: extract key features from the analysis data, such as optical power fluctuation amplitude, fluctuation frequency, fluctuation period, fluctuation mutation point, etc., to construct a multi-dimensional feature space.

[0044] Model training uses machine learning algorithms, such as support vector machines, decision trees, neural networks, etc., to train historical data and establish a fault prediction model. Feature engineering is also used to improve the training effect of the model.

[0045] Model validation: The model is validated through cross-validation, leave-one-out method and other methods to evaluate its prediction accuracy and robustness. And the confusion matrix, ROC curve and other evaluation indicators are used to quantify the model performance.

[0046] Model optimization: continuously optimize the fault prediction model, and use ensemble learning, hyperparameter tuning and other technologies to improve the model's generalization ability and prediction accuracy.

[0047] S5: Intelligent early warning, including real-time data collection, real-time prediction, multi-level early warning mechanism, automatic dispatch and closed-loop feedback.

[0048] Real-time data collection, constantly collect new data from the network, pre-process and store it. Use distributed data storage and processing technology to ensure the efficiency and scalability of the system.

[0049] Real-time prediction: input new data into the fault prediction model, make real-time predictions, and output warning results. The real-time prediction system can quickly respond to sudden changes in optical power and provide timely fault warnings.

[0050] Multi-level warning mechanism: According to the severity of the warning results, a multi-level warning mechanism is set up, from general warning to emergency warning, and different treatment measures are taken respectively. The multi-level warning mechanism can effectively distinguish different degrees of optical power fluctuations and improve processing efficiency.

[0051] Automatic dispatching: Based on the early warning results, the artificial intelligence system automatically generates work orders and dispatches them to relevant maintenance personnel to ensure timely processing. The work order system records the detailed process and results of each fault handling, providing data support for subsequent analysis.

[0052] Closed-loop feedback: After the maintenance personnel have handled the fault, they will feed back the processing results to the system. The system will optimize the fault prediction model based on the feedback results to achieve closed-loop management of fault handling.

[0053] S6: Data visualization and report generation, including real-time monitoring interface, historical data query and automatic report generation.

[0054] The real-time monitoring interface provides an intuitive monitoring interface that displays optical power data, warning information, and processing status in real time, helping operation and maintenance personnel to quickly understand network conditions.

[0055] Historical data query: supports historical data query and analysis, helping operation and maintenance personnel to review and analyze past optical power fluctuations and optimize fault handling strategies.

[0056] Automatic report generation: The system automatically generates an optical power fluctuation analysis report, which includes data statistics, anomaly detection, fault prediction and processing results, providing decision support for management.

[0057] Based on the above solution, the key nodes of the process of a method for intelligent early warning of optical power fluctuation based on broadband fixed-line FTTH in this embodiment are:

[0058] 1. Configure the data collection strategy for optical port devices, define the basic information such as optical power value, device ID, timestamp, etc., to ensure that the data fully covers each optical port device in the network.

[0059] 2. Collect optical power data from each optical port device in real time through the SNMP protocol to ensure that the collection frequency is reasonable so as to accurately capture optical power fluctuations.

[0060] 3. Collect users’ offline records from the business system and analyze data that may be related to optical power fluctuations. Integrate user device usage habit data to form a complete data view.

[0061] 4. Preprocess the collected optical power data, including data cleaning, format conversion and data standardization. Delete duplicate data, correct erroneous data, fill in missing data, and convert data from different sources into a unified format to eliminate differences between different devices.

[0062] 5. Using big data analysis technology, comprehensively analyze OLT equipment data, user offline data and probe detection data, extract the key features of optical power fluctuations, identify periodic changes and sudden anomalies in the data, and obtain the distribution range of the optical power values ​​of each optical port.

[0063] 6. Use anomaly detection algorithms to identify optical power values ​​that are beyond the normal range and determine potential faulty optical ports. Use algorithms such as isolation forest and support vector machines to improve the accuracy and reliability of anomaly detection.

[0064] 7. Based on the data analysis results, a fault prediction model for optical power fluctuation is constructed. The model is trained through machine learning algorithms to extract features such as optical power fluctuation amplitude and fluctuation frequency, and the prediction accuracy and robustness of the model are evaluated through methods such as cross-validation.

[0065] 8. Deploy the trained fault prediction model to the online system to process and analyze the newly collected optical power data in real time. Output warning information based on the real-time prediction results, and classify and process warnings of different severity through a multi-level warning mechanism.

[0066] 9. Based on the early warning results, the artificial intelligence system automatically generates work orders and sends them to relevant maintenance personnel to ensure that the fault can be handled in a timely manner. After the maintenance personnel handle the fault, they will feed back the processing results to the system to optimize the fault prediction model and achieve closed-loop management.

[0067] In summary, the present invention realizes intelligent monitoring of optical power fluctuations. Through big data analysis and artificial intelligence technology, it monitors and analyzes optical power fluctuations in real time, promptly discovers potential faults, automates fault warnings and dispatches, automatically generates fault warnings and dispatches them to maintenance personnel, improves fault handling efficiency, reduces operation and maintenance costs, and solves the problems of untimely monitoring and low processing efficiency for optical power fluctuations based on broadband fixed-line FTTH in the prior art.

[0068] This embodiment also discloses a device for intelligent early warning of optical power fluctuation based on broadband fixed-line FTTH network.

[0069] Reference Figure 2 The device based on the intelligent early warning of optical power fluctuation of broadband fixed network FTTH includes a data acquisition module, a data preprocessing module, a data analysis module, a fault prediction module, an intelligent early warning module and a data visualization and report generation module connected in sequence. Among them, the data acquisition module is used to obtain data such as optical power data, device ID and timestamp. The data preprocessing module is used for data cleaning, format conversion, standardization and enhancement. The data analysis module is used for data mining, statistical analysis, anomaly detection and trend analysis. The fault prediction module is used for feature extraction, model training, verification and optimization. The intelligent early warning module is used for real-time data acquisition, real-time prediction, multi-level early warning mechanism, automatic dispatching and closed-loop feedback. The data visualization and report generation module is used to form a real-time monitoring interface, historical data query and automatic report generation. Through the collaborative work of the above modules, the whole process from data acquisition to fault prediction and intelligent early warning can be effectively realized, providing efficient early warning support for network operation and maintenance personnel, and ensuring the stable and efficient operation of the optical fiber network.

[0070] The above are all preferred embodiments of the present invention, and the protection scope of the present application is not limited thereto. Therefore, any equivalent changes made according to the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. An intelligent early warning method for optical power fluctuation based on broadband fixed-line FTTH network, characterized in that: The following steps are involved: S1: Data collection, including collecting the optical power value, device ID and timestamp of the optical port device; S2: Data preprocessing, including data cleaning, format conversion, data standardization and data enhancement; S3: Data analysis, including data mining, statistical analysis, anomaly detection, and trend analysis; S4: Fault prediction, including feature extraction, model training, model validation, and model optimization; S5: Intelligent early warning, including real-time data collection, real-time prediction, multi-level early warning mechanism, automatic dispatching and closed-loop feedback; S6: Data visualization and report generation, including real-time monitoring interface, historical data query and automatic report generation.

2. According to claim 1, the method for intelligent early warning of optical power fluctuation based on broadband fixed-line FTTH network is characterized in that: In S1, the optical power value, device ID and timestamp are collected from each optical port device in the network through the SNMP protocol, including PON optical power collection and optical modem optical power collection; the user's offline records are collected from the business system, and the data related to optical power fluctuations are analyzed from them, and the user's device usage habits are integrated to form a complete data view.

3. According to claim 1, the intelligent early warning method for optical power fluctuation based on broadband fixed-line FTTH network is characterized in that: In S2, data cleaning includes: cleaning the collected data, including deleting duplicate data, correcting erroneous data, filling missing data, ensuring the integrity and accuracy of the data, and identifying and correcting outliers through abnormal data algorithms, such as Z-score, IQR and Boxpiot; format conversion includes: converting data from different sources into a unified format and using standardized format data; data standardization includes: standardizing the optical power value, using normalization or Z-score standardization methods, and unifying the measurement units of the data; data enhancement includes: enhancing the data through interpolation and smoothing methods.

4. According to claim 1, the intelligent early warning method for optical power fluctuation based on broadband fixed-line FTTH network is characterized in that: In S3, data mining includes: using data mining algorithms, such as classification algorithms, clustering algorithms, and regression algorithms, to extract the changing trends and abnormal fluctuations of optical power values ​​from a large amount of data, and identifying different types of optical power fluctuation patterns through clustering analysis; statistical analysis includes: performing statistical analysis on optical power values, calculating mean, variance, and standard deviation indicators, determining the range of normal optical power values, and identifying the normal fluctuation range and abnormal fluctuation of optical power through probability distribution analysis; anomaly detection includes: using anomaly detection algorithms such as isolation forests and support vector machines to identify optical power values ​​that exceed the normal range and determine potential faulty optical ports; trend analysis includes: performing time series analysis on optical power data, identifying long-term trends and periodic fluctuations, and predicting future optical power change trends.

5. According to claim 1, the intelligent early warning method for optical power fluctuation based on broadband fixed-line FTTH network is characterized by: In S4, feature extraction includes: extracting key features from the analysis data, such as optical power fluctuation amplitude, fluctuation frequency, fluctuation period, and fluctuation mutation point, and constructing a multi-dimensional feature space; model training includes: using machine learning algorithms, such as support vector machines, decision trees, and neural networks, to train historical data, establish a fault prediction model, and improve the training effect of the model through feature engineering; model verification includes: verifying the model through cross-validation and leave-one-out method, evaluating its prediction accuracy and robustness, and using confusion matrix and ROC curve evaluation indicators to quantify model performance; model optimization includes: optimizing the fault prediction model, using ensemble learning and hyperparameter tuning techniques to improve the generalization ability and prediction accuracy of the model.

6. According to claim 1, the intelligent early warning method for optical power fluctuation based on broadband fixed-line FTTH network is characterized by: In S5, real-time data collection includes: continuously collecting new data in the network, using distributed data storage and processing technology for preprocessing and storage; real-time prediction includes: inputting new data into the fault prediction model, performing real-time prediction, outputting warning results, and real-time predicting the system's response to sudden changes in optical power to provide fault warnings; multi-level warning mechanism includes: setting a multi-level warning mechanism according to the severity of the warning results, taking different treatment measures from general warnings to emergency warnings, and using a multi-level warning mechanism to distinguish different degrees of optical power fluctuations; automatic dispatching includes: automatically generating work orders through the artificial intelligence system based on the warning results, and dispatching them to relevant maintenance personnel. The work order system records the detailed process and results of each fault handling to provide data support for subsequent analysis; closed-loop feedback includes: after the maintenance personnel have handled the fault, they will feed back the processing results to the system. The system optimizes the fault prediction model based on the feedback results to achieve closed-loop management of fault handling.

7. According to claim 1, the intelligent early warning method for optical power fluctuation based on broadband fixed-line FTTH network is characterized by: In S6, the real-time monitoring interface includes: providing a monitoring interface to display optical power data, warning information and processing status in real time; historical data query includes: supporting historical data query and analysis, and optimizing fault handling strategies; automatic report generation includes: the system automatically generates an optical power fluctuation analysis report, including data statistics, anomaly detection, fault prediction and processing results.

8. A device for intelligent early warning of optical power fluctuation based on broadband fixed-line FTTH network, used in the intelligent early warning method for optical power fluctuation based on broadband fixed-line FTTH network as claimed in any one of claims 1 to 7, characterized in that: It includes a data acquisition module, a data preprocessing module, a data analysis module, a fault prediction module, an intelligent early warning module and a data visualization and report generation module which are connected in sequence.

Images