Optical fiber network hidden trouble type intelligent identification method, system, device and medium

Abstract

The present disclosure provides a kind of optical fiber network hidden danger type intelligent identification method and system, equipment, medium, the method comprises the following steps: S1, obtains the historical data related to ONU offline history alarm event.According to the training of classification model, hidden danger judgment model is obtained.S2, in the case where ONU offline alarm event occurs, the real-time data related to current ONU offline alarm event is obtained.S3, according to the real-time data related to current ONU offline alarm event, the probability that current ONU offline alarm event is hidden danger is obtained by using hidden danger judgment model.S4, according to the probability that current ONU offline alarm event is hidden danger and preset threshold value determines whether current ONU offline alarm event is hidden danger according to current ONU offline alarm event.If yes, according to preset rule, the hidden danger type of current ONU offline alarm event is determined.The present disclosure can accurately distinguish indoor optical fiber hidden danger and primary / secondary optical splitter hidden danger, realize the high-precision automatic identification of two typical optical fiber link hidden dangers.

Description

Technical Field

[0001] This disclosure relates to the field of intelligent identification technology for fiber optic network vulnerability types, and in particular to an intelligent identification method and system for fiber optic network vulnerability types, electronic equipment, computer-readable storage medium, and computer program product. Background Technology

[0002] In the context of large-scale deployment of fiber optic access networks (especially Passive Optical Networks, or PONs), the online stability of the Optical Network Unit (ONU), as the user-side terminal equipment, directly affects the quality of broadband service. However, in actual operation and maintenance, frequent ONU offline is a common and frequent fault phenomenon. Its causes are complex, and may stem from physical link problems such as aging, bending, and connector contamination of the user-side fiber, or from potential problems at the central office end such as loose ports of the primary or secondary optical splitter, changes in environmental temperature and humidity, and aging of components. Failure to quickly and accurately identify the root cause of the fault will lead to wasted operation and maintenance resources, decreased user satisfaction, and even a batch of complaints.

[0003] Currently, the industry mainly relies on the following two technical solutions to handle ONU offline alarms: 1. Alarm filtering mechanism based on a single threshold or frequency Existing network management systems typically set an "ONU offline count threshold" (e.g., more than 5 times a day) as the condition for triggering work orders. While this method is simple to implement, it lacks comprehensive analysis of the alarm context and cannot distinguish between user-initiated power outages, hotel converged terminal service characteristics, centralized power outages in commercial buildings, and actual fiber optic link vulnerabilities. For example, hotel-customized ONUs may periodically restart due to business strategies, generating numerous "pseudo-offline" alarms; unified power outages in urban commercial buildings during non-business hours can also cause a batch of ONUs to go offline. If these scenarios are misjudged as network failures, it will result in a large number of invalid work orders.

[0004] 2. A batch fault identification method based on topological clustering

[0005] Some advanced operators have attempted to introduce the "offline clustering" of users connected to optical splitters as a criterion for judgment. That is, when more than a certain proportion of users connected to a certain optical splitter go offline at the same time, the splitter is judged to have a potential problem. This method has a certain effect in identifying optical splitter-level faults, but it has obvious limitations: it does not consider the time window of alarm occurrence, and the clustered offline at night or on holidays may be due to external power supply rather than network problems; it does not combine optical power data, so it cannot distinguish whether it is a problem with the optical splitter or the fiber optic cable at the drop end; it lacks the ability to effectively identify scenarios where a single user is offline at high frequency but without clustering (a typical potential problem at the drop end); and it does not exclude interference from specific equipment types or cell types, resulting in a high false alarm rate.

[0006] Furthermore, although recent studies have attempted to introduce machine learning models to classify the causes of ONU offline operations, these methods have not yet been widely implemented in live networks due to issues such as high costs of training sample labeling, reliance on expert experience for feature engineering, and poor model interpretability. Most solutions remain in the laboratory stage and are unable to meet the stringent requirements of telecom-grade operations and maintenance for accuracy, real-time performance, and traceability.

[0007] In summary, existing technologies generally suffer from the following drawbacks: (1) Alarm judgment dimension is single and lacks fusion analysis of multi-source heterogeneous features: Existing solutions mostly rely on offline number or simple aggregation threshold for judgment, without comprehensively considering multi-dimensional information such as alarm occurrence time, user cell type, ONU equipment model, optical power degradation degree and splitter topology, which makes it impossible to accurately distinguish between real network hidden dangers and external interference events.

[0008] (2) Failure to effectively eliminate known interference factors such as commercial power outages and hotel terminals: For unified power outages in urban commercial buildings, planned power outages in rural areas, and business scenarios such as hotels using specific converged terminals, the existing methods have failed to establish targeted exclusion rules, resulting in a large number of non-fault alarms being misjudged as network vulnerabilities and generating invalid work orders.

[0009] (3) It is impossible to identify two typical problems, namely "downlink fiber optic problems" and "optical splitter problems", under the same framework: downlink fiber optic problems usually manifest as high-frequency offline of a single user and severe optical attenuation, but without spatial clustering; while first- or second-level optical splitter problems manifest as multiple users going offline in a concentrated manner under the splitter, which is recoverable (the duration is within a certain range), but the optical attenuation may not be degraded. Existing technology lacks a unified discrimination logic that covers both types of fault modes at the same time, resulting in one type of problem being missed and the other being misjudged.

[0010] (4) The rule logic is disconnected from the business scenario, resulting in a high false alarm / false alarm rate: Although some solutions introduce the concept of clustering, they do not limit the effective time window (e.g., only offline during working hours has diagnostic value) and do not combine the duration of a single offline event with light decay data for cross-validation, resulting in insufficient reliability of the judgment results. Summary of the Invention

[0011] This disclosure provides a method and system for intelligent identification of fiber optic network vulnerability types, an electronic device, a computer-readable storage medium, and a computer program product.

[0012] Firstly, this disclosure provides an intelligent identification method for fiber optic network vulnerability types. This method includes: S1, acquiring historical data related to historical alarm events related to ONU offline events. A classification model is trained based on the historical data to obtain a vulnerability judgment model. S2, real-time monitoring of the ONU's status; in the event of an ONU offline alarm event, acquiring real-time data related to the current ONU offline alarm event. S3, based on the real-time data related to the current ONU offline alarm event, using the vulnerability judgment model to obtain the probability that the current ONU offline alarm event is a vulnerability. S4, determining whether the current ONU offline alarm event is a vulnerability based on the probability of it being a vulnerability and a preset threshold. If so, determining the vulnerability type of the current ONU offline alarm event according to preset rules.

[0013] Optionally, S1 includes: S11, collecting historical alarm events related to multiple ONU offline events within a preset historical time period, and obtaining historical data related to the historical alarm events related to multiple ONU offline events and the potential hazard type marked by the operation and maintenance expert for each historical alarm event related to ONU offline events. The marked potential hazard types include: fiber optic hazard, optical splitter hazard, or no hazard. The historical data related to the historical alarm events related to multiple ONU offline events includes: whether the current alarm event is an ONU offline alarm, whether the current alarm event occurred during working hours, whether the community where the user involved in the current alarm event belongs is an urban or rural residential community, whether the ONU model involved in the current alarm event belongs to an exclusion list, the number of times the ONU involved in the current alarm event went offline on the same day, the aggregation degree of the first-level / second-level optical splitter to which the ONU involved in the current alarm event belongs, the single offline time in the current alarm event, and the optical attenuation value most recently reported by the ONU related to the current alarm event. The clustering degree of the primary / secondary optical splitters to which the ONUs involved in this alarm event belong is the maximum of the percentages of users whose primary optical splitters to which the ONUs involved in this alarm event have gone offline more than 3 times and the percentages of users whose secondary optical splitters to which the ONUs involved in this alarm event have gone offline more than 3 times. S12: Construct historical feature vectors based on historical data related to historical alarm events involving multiple ONU offline events. Combine each historical feature vector with the hazard type labeled by the corresponding operation and maintenance experts for each historical alarm event involving ONU offline events to obtain a feature dataset. S13: Train the classification model based on the feature dataset to obtain a hazard judgment model.

[0014] Optionally, the historical feature vector includes feature values ​​in eight dimensions. Specifically: In the first dimension, if the current alarm event is an ONU offline alarm, the feature value is 1; otherwise, the feature value is 0. In the second dimension, if the current alarm event occurs during working hours, the feature value is 1; otherwise, the feature value is 0. In the third dimension, if the user involved in the current alarm event belongs to an urban or rural residential area, the feature value is 1; otherwise, the feature value is 0. In the fourth dimension, if the ONU model involved in the current alarm event is not in the exclusion list, the feature value is 1; otherwise, the feature value is 0. In the fifth dimension, the feature value equals the number of times the ONU involved in the current alarm event has gone offline that day. In the sixth dimension, the feature value equals the aggregation degree of the first-level / second-level splitter to which the ONU involved in the current alarm event belongs. In the seventh dimension, the feature value equals the single offline time in the current alarm event. In the eighth dimension, the feature value equals the optical attenuation value most recently reported by the ONU related to the current alarm event.

[0015] Optionally, S3 includes: S31, extracting features from the real-time data related to the current ONU offline alarm event to obtain a feature vector of the real-time data related to the current ONU offline alarm event; and S32, inputting the feature vector of the real-time data related to the current ONU offline alarm event into the hazard judgment model. The output of the hazard judgment model is the probability that the current ONU offline alarm event is a hazard.

[0016] Optionally, S4 includes: S41, comparing the probability that the current ONU offline alarm event is a potential hazard with a corresponding preset threshold to determine whether the current ONU offline alarm event is a potential hazard. S42, determining the hazard type of the current ONU offline alarm event based on whether the current ONU offline alarm event is a potential hazard and preset rules.

[0017] Optionally, S42 includes: S421, determining whether the features in the feature vector of the real-time data related to the current ONU offline alarm event meet a first preset condition. If yes, then execute S422; otherwise, execute S424. The first preset condition includes: the current alarm event is an ONU offline alarm; the current alarm event occurs during working hours; the user involved in the current alarm event belongs to an urban or rural residential community; the ONU model involved in the current alarm event is not a model in the exclusion list; and the number of times the ONU involved in the current alarm event has gone offline on the same day is greater than a preset number of offline events. S422, determining whether the features in the feature vector of the real-time data related to the current ONU offline alarm event meet a second preset condition. If yes, then determine that the hidden danger type of the current ONU offline alarm event is a fiber optic hidden danger. Otherwise, execute S423. The second preset condition includes: the optical attenuation value most recently reported by the ONU related to the current alarm event is less than a preset optical attenuation value. S423, determining whether the features in the feature vector of the real-time data related to the current ONU offline alarm event meet a third preset condition. If so, determine the current ONU offline alarm event's vulnerability type as an optical splitter vulnerability. Otherwise, execute S424. The third preset conditions include: the aggregation degree of the primary / secondary optical splitter to which the ONU involved in this alarm event belongs is greater than the preset aggregation degree, and the total number of users connected to the ONU involved in this alarm event is greater than or equal to the preset number of users, and the single offline time in this alarm event is within the preset offline period. S424, generate a work order and push it to the management personnel for manual verification.

[0018] Secondly, this disclosure provides an intelligent identification system for fiber optic network vulnerability types, comprising: a model acquisition module, a real-time data acquisition module, a probability determination module, and a vulnerability type determination module. The model acquisition module is configured to: acquire historical data related to historical alarm events related to ONU offline events; train a classification model based on the historical data to obtain a vulnerability judgment model. The real-time data acquisition module is configured to: monitor the ONU status in real time; and acquire real-time data related to the current ONU offline alarm event when an ONU offline alarm event occurs. The probability determination module is configured to: use the vulnerability judgment model to determine the probability that the current ONU offline alarm event is a vulnerability based on the real-time data related to the current ONU offline alarm event. The vulnerability type determination module is configured to: determine whether the current ONU offline alarm event is a vulnerability based on the probability that it is a vulnerability and a preset threshold; if so, determine the vulnerability type of the current ONU offline alarm event according to preset rules.

[0019] Thirdly, this disclosure provides an electronic device comprising: a plurality of processing cores; and an on-chip network configured to interact with data between the plurality of processing cores and external data; wherein one or more of the processing cores store one or more instructions, and the one or more instructions are executed by the one or more processing cores to enable the one or more processing cores to perform the above-described intelligent identification method for fiber optic network vulnerability types.

[0020] Fourthly, this disclosure provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the above-mentioned intelligent identification method for fiber optic network vulnerability types.

[0021] Fifthly, this disclosure provides a computer program product that includes computer-readable code or a non-volatile computer-readable storage medium carrying computer-readable code. When the computer-readable code is run in a processor of an electronic device, the processor in the electronic device executes the above-described intelligent identification method for fiber optic network vulnerability types.

[0022] The embodiments provided in this disclosure employ a two-tier architecture of "initial screening via machine learning + refined judgment based on rules." The final determination is based on combined judgment rules with business semantics, resulting in logically clear, interpretable, and traceable outcomes that meet telecom-grade operation and maintenance requirements. With the assistance of machine learning models, high-precision, low-false-report, and interpretable automatic identification of two typical fiber optic link vulnerabilities is achieved, improving PON network operation and maintenance efficiency and fault location accuracy.

[0023] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0024] The accompanying drawings are provided to further illustrate the present disclosure and form part of the specification. They are used together with the embodiments of the present disclosure to explain the disclosure and do not constitute a limitation thereof. The above and other features and advantages will become more apparent to those skilled in the art from the detailed description of exemplary embodiments with reference to the accompanying drawings, in which: Figure 1 A flowchart illustrating an intelligent identification method for fiber optic network vulnerability types provided in this embodiment of the disclosure; Figure 2 A flowchart illustrating another intelligent identification method for fiber optic network vulnerability types provided in this embodiment of the disclosure; Figure 3 A flowchart of another intelligent identification method for fiber optic network vulnerability types provided in this embodiment of the disclosure; Figure 4 A flowchart of another intelligent identification method for fiber optic network vulnerability types provided in this embodiment of the disclosure; Figure 5 A flowchart of another intelligent identification method for fiber optic network vulnerability types provided in this embodiment of the disclosure; Figure 6 A block diagram of an intelligent identification system for fiber optic network vulnerability types provided in this embodiment of the disclosure; Figure 7 This is a block diagram of an electronic device provided in an embodiment of the present disclosure. Detailed Implementation

[0025] To enable those skilled in the art to better understand the technical solutions of this disclosure, exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments of this disclosure to aid understanding. These should be considered merely exemplary. Therefore, those skilled in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0026] Where there is no conflict, the various embodiments of this disclosure and the features thereof in the embodiments may be combined with each other.

[0027] As used herein, the term “and / or” includes any and all combinations of one or more related enumerated entries.

[0028] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit this disclosure. As used herein, the singular forms “a” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will also be understood that when the terms “comprising” and / or “made of” are used in this specification, the presence of the stated feature, integral, step, operation, element, and / or component is specified, but the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or groups thereof is not excluded. Words such as “connected” or “linked” are not limited to physical or mechanical connections but can include electrical connections, whether direct or indirect.

[0029] Unless otherwise specified, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art. It will also be understood that terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant art and this disclosure, and will not be interpreted as having an idealized or overly formal meaning, unless expressly so defined herein.

[0030] The intelligent identification method for fiber optic network vulnerability types according to embodiments of this disclosure can be executed by electronic devices such as terminal devices or servers. Terminal devices can be in-vehicle devices, user equipment (UE), mobile devices, user terminals, terminals, cellular phones, cordless phones, personal digital assistants (PDAs), handheld devices, computing devices, wearable devices, etc. The method can be implemented by a processor calling computer-readable program instructions stored in memory. Alternatively, the method can be executed by a server.

[0031] Embodiments of this disclosure provide an intelligent identification method for types of hidden dangers in fiber optic networks, such as... Figure 1 As shown, it includes steps S1 to S4.

[0032] Step S1: Obtain historical data related to offline alarm events of the ONU. Train the classification model based on the historical data to obtain the hazard judgment model.

[0033] Step S2: Monitor the status of the ONU in real time, and in the event of an ONU offline alarm event, obtain real-time data related to the current ONU offline alarm event.

[0034] Step S3: Based on the real-time data related to the current ONU offline alarm event, use the hidden danger judgment model to obtain the probability that the current ONU offline alarm event is a hidden danger.

[0035] Step S4: Determine whether the current ONU offline alarm event is a potential hazard based on the probability that it is a hidden danger and a preset threshold. If so, determine the hazard type of the current ONU offline alarm event according to preset rules.

[0036] This disclosure provides an intelligent identification method for fiber optic network vulnerability types, used to automatically distinguish between fiber optic network vulnerabilities and optical splitter vulnerabilities in broadband access systems. It combines a supervised learning model with an expert rule engine, using machine learning algorithms to perform preliminary classification based on historical data, and then using high-precision business rules for final determination, balancing model generalization ability and operational interpretability.

[0037] In some embodiments, as shown in FIG2, step S1 includes: Step S11: Collect historical alarm events of multiple ONU offline events within a preset historical time period, and obtain historical data related to the historical alarm events of multiple ONU offline events and the potential hazard types marked by operation and maintenance experts for each historical alarm event of ONU offline events.

[0038] In step S11, the identified hazard types include: fiber optic cable hazard, optical splitter hazard, or no hazard. Historical data related to multiple ONU offline alarm events includes: whether the current alarm event is an ONU offline alarm, whether the current alarm event occurred during working hours, whether the user involved in the current alarm event resides in an urban or rural residential area, whether the ONU model involved in the current alarm event is on the exclusion list, the number of times the ONU involved in the current alarm event went offline that day, the aggregation degree of the primary / secondary splitter to which the ONU involved in the current alarm event belongs, the duration of a single offline event in the current alarm event, and the optical attenuation value most recently reported by the ONU related to the current alarm event. The aggregation degree of the primary / secondary splitter to which the ONU involved in the current alarm event belongs is the maximum of the percentage of users whose primary splitter to which the ONU involved in the current alarm event has experienced more than 3 offline events and the percentage of users whose secondary splitter to which the ONU involved in the current alarm event has experienced more than 3 offline events.

[0039] For example, the preset historical period can refer to the most recent 6-12 months. Multiple ONU offline historical alarm events within the preset historical period can be collected from the network management system, and / or resource management system, and / or performance monitoring platform. Historical data related to multiple ONU offline historical alarm events may include ONU offline alarm records, user attributes (cell type, ONU model), optical attenuation values, splitter topology relationships, alarm duration, and other data.

[0040] For example, the models excluded from the list include those with frequent offline service characteristics, such as 'Fiberhome MR222 (Ethernet Passive Optical Network, EPON) (converged terminal)', 'MR222-K converged terminal', 'Fiberhome MR820 (Local Area Network, LAN) (converged terminal)', 'SK-M524 (dual-band PON converged terminal)', 'DT541-PON converged terminal', and other converged terminals customized by WoHotel.

[0041] Understandably, the method for determining whether the ONU model involved in this alarm event belongs to the exclusion list introduces an explicit business interference exclusion mechanism. This method excludes the impact of commercial power outages and avoids periodic offline interference from customized terminals such as those used by Wo Hotel. This exclusion mechanism, based on a deep understanding of the existing network business scenarios, significantly reduces the false alarm rate and is a key technical means to improve system usability.

[0042] By using feature 4 to exclude the impact of commercial power outages and by using feature 5 to avoid periodic offline interference from customized terminals such as those used by Wo Hotel, this exclusion mechanism, based on a deep understanding of existing network business scenarios, significantly reduces the false alarm rate and is a key technical means to improve the system's usability.

[0043] For example, the calculation method for the aggregation degree of the primary / secondary optical splitter to which the ONU involved in this alarm event belongs is as follows: For the primary optical splitter, if the total number of users connected to it is N1, and the number of users who have ≥3 ONU offline alarms on the same day is M1, the aggregation degree of the primary optical splitter is calculated as M1 / ​​N1. For the secondary optical splitter, if the total number of users connected to it is N2, and the number of users who have ≥3 ONU offline alarms on the same day is M2, the aggregation degree of the secondary optical splitter is M2 / N2. The aggregation degree of the primary / secondary optical splitter to which the ONU involved in this alarm event belongs is the maximum value between M1 / ​​N1 and M2 / N2. The calculation method for the single offline time in this alarm event can be: calculate the value of Δt, Δt = recovery time - alarm generation time (unit: minutes). It can be understood that only when Δt∈[1, 120] minutes is the offline event considered recoverable, which meets the time characteristics of typical hidden dangers such as poor contact of optical splitters.

[0044] Step S12: Construct historical feature vectors based on historical data related to historical alarm events that are offline for multiple ONUs. Combine each historical feature vector with the potential hazard type marked by the corresponding operation and maintenance expert for each offline historical alarm event of ONU to obtain a feature dataset.

[0045] Understandably, after obtaining historical data related to offline historical alarm events of multiple ONUs, it is advisable to perform cleaning, normalization, and other processing before feature extraction.

[0046] Step S13: Train the classification model based on the feature dataset to obtain the hidden danger judgment model.

[0047] For example, the classification model can be a lightweight model, such as logistic regression, decision trees, or extreme gradient boosting (XGBoost), whose output is the probability of a hazard or a candidate label. A binary classification model is trained using interpretable machine learning algorithms (such as decision trees or logistic regression) with the goal of predicting "whether it belongs to one of two types of hazards". The model output may include: hazard probability scores and feature importance rankings (used to verify the rationality of the business logic).

[0048] In some embodiments, the historical feature vector includes feature values ​​in eight dimensions. Specifically: In the first dimension, if the current alarm event is an ONU offline alarm, the feature value is 1; otherwise, the feature value is 0. In the second dimension, if the current alarm event occurs during working hours, the feature value is 1; otherwise, the feature value is 0. In the third dimension, if the user involved in the current alarm event belongs to an urban or rural residential community, the feature value is 1; otherwise, the feature value is 0. In the fourth dimension, if the ONU model involved in the current alarm event is not in the exclusion list, the feature value is 1; otherwise, the feature value is 0. In the fifth dimension, the feature value equals the number of times the ONU involved in the current alarm event has gone offline that day. In the sixth dimension, the feature value equals the aggregation degree of the first-level / second-level splitter to which the ONU involved in the current alarm event belongs. In the seventh dimension, the feature value equals the single offline time in the current alarm event. In the eighth dimension, the feature value equals the optical attenuation value most recently reported by the ONU related to the current alarm event.

[0049] Understandably, the above methods can construct a multi-dimensional feature combination system encompassing time, scenario, device, performance, and topology. This system integrates features such as time window, cell type, model excluding specific hotel converged terminals, daily offline frequency per user, number and clustering of users connected to the splitter, duration of single offline events, and optical attenuation to form a structured discrimination basis. This feature combination system accurately models typical interference sources and fault modes in actual operation and maintenance, forming the foundation for achieving high-accuracy hazard identification.

[0050] In some embodiments, such as Figure 3 As shown, step S3 includes steps S31 and S32.

[0051] Step S31: Extract features from the real-time data related to the current ONU offline alarm event to obtain the feature vector of the real-time data related to the current ONU offline alarm event.

[0052] For example, an 8-dimensional vector can be extracted using the method described above.

[0053] Step S32: Input the feature vector of real-time data related to the current ONU offline alarm event into the hidden danger judgment model.

[0054] In step S32, the output of the hidden danger judgment model is the probability that the current ONU offline alarm event is a hidden danger.

[0055] Understandably, classification models can determine whether an alarm event is a potential hazard or not. They do not directly output the final category but are only used to narrow down the candidate range, thus avoiding false alarms.

[0056] In some embodiments, such as Figure 4 As shown, step S4 includes steps S41 and S42.

[0057] Step S41: Compare the probability that the current ONU offline alarm event is a potential hazard with the corresponding preset threshold to determine whether the current ONU offline alarm event is a potential hazard.

[0058] For example, the feature vector is input into the trained model. If the model outputs that the probability of the current ONU offline alarm event being a potential hazard is greater than or equal to a preset threshold (such as 0.6), then proceed to S42. Otherwise, directly determine that the current ONU offline alarm event is not a potential hazard and end the process.

[0059] Step S42: Determine the potential risk type of the current ONU offline alarm event based on whether the current ONU offline alarm event is a potential risk and the preset rules.

[0060] Understandably, accurate judgments can be made using the following preset rules.

[0061] In some embodiments, such as Figure 5 As shown, step S42 includes steps S421 to S424.

[0062] Step S421: Determine whether the features in the feature vector of the real-time data related to the current ONU offline alarm event meet the first preset condition. If yes, proceed to step S422; otherwise, proceed to step S424.

[0063] In step S421, the first preset conditions include: the alarm event is an ONU offline alarm, the alarm event occurs during working hours, the user involved in the alarm event belongs to an urban or rural residential community, the ONU model involved in the alarm event is not in the exclusion list, and the number of times the ONU involved in the alarm event goes offline on the same day is greater than the preset number of offline events.

[0064] For example, if the preset number of offline events is 3, the first preset condition is met if the ONU involved in this alarm event is offline, occurs during working hours, the user's community is a residential community, the ONU model involved in this alarm event is a model not included in the exclusion list, and the ONU involved in this alarm event has more than 3 offline events on the same day.

[0065] Step S422: Determine whether the features in the feature vector of the real-time data related to the current ONU offline alarm event meet the second preset condition. If yes, determine the vulnerability type of the current ONU offline alarm event as a fiber optic vulnerability. Otherwise, proceed to S423.

[0066] In step S422, the second preset condition includes: the optical attenuation value most recently reported by the ONU related to this alarm event is less than the preset optical attenuation value.

[0067] For example, if the preset optical attenuation value is -27 dB, then the following conditions must be met: the ONU involved in this alarm event is offline; it occurs during working hours; the user's community is a residential community; the ONU model involved in this alarm event is a model not included in the exclusion list; the ONU involved in this alarm event has been offline more than 3 times on the same day; and the optical attenuation value most recently reported by the ONU related to this alarm event is less than -27 dB. In this case, the potential risk type of the current ONU offline alarm event is considered to be a potential fiber optic vulnerability.

[0068] Step S423: Determine whether the features in the feature vector of the real-time data related to the current ONU offline alarm event meet the third preset condition. If yes, determine that the potential hazard type of the current ONU offline alarm event is an optical splitter hazard. Otherwise, proceed to S424.

[0069] In step S423, the third preset conditions include: the aggregation degree of the primary / secondary optical splitter to which the ONU involved in this alarm event belongs is greater than the preset aggregation degree, and the total number of users connected to the ONU involved in this alarm event is greater than or equal to the preset number of users, and the single offline time in this alarm event is within the preset offline time period.

[0070] For example, if the preset aggregation degree is 50%, the preset number of users is 4, and the preset offline period is 120 minutes, then if the ONU involved in this alarm event is offline, occurs during working hours, the designed user belongs to a residential community, the ONU model involved in this alarm event is a model not included in the exclusion list, the ONU involved in this alarm event has more than 3 offline occurrences on the same day, and the aggregation degree of the primary / secondary optical splitter to which the ONU involved in this alarm event belongs is greater than 50%, the total number of users connected to the ONU involved in this alarm event is greater than or equal to 4, and the single offline time in this alarm event is between [1, 120], then the hidden danger type of the current ONU offline alarm event is determined to be an optical splitter hidden danger.

[0071] Understandably, the third precondition not only calculates the clustering degree but also explicitly requires that the object being judged must belong to a primary or secondary optical splitter and have at least four connected users, thus avoiding misjudgments due to accidental offline events in small-scale optical splitting scenarios. This design enhances the applicability and robustness of the clustering degree metric in real network topologies.

[0072] Step S424: Generate a work order and push it to the administrator for manual verification.

[0073] Understandably, if only the first preset condition is met but the second and third preset conditions are not met, or in other words, the first preset condition is not met, manual verification is required. Understandably, the above rule logic is inherited from the experience of operations and maintenance experts, has strong business constraints, and ensures that the final result is explainable and traceable.

[0074] Through steps S421 to S424, the embodiments of this disclosure design differentiated judgment logic rules for two types of potential hazards: for fiber optic malfunctions, a combination of criteria including high-frequency offline detection, optical attenuation, and lack of splitter aggregation is used. For malfunctions of primary or secondary optical splitters, high-frequency offline detection, splitter aggregation, and offline recoverability are used. This dual-path judgment logic can effectively distinguish between two types of hazards with drastically different physical causes within a unified framework, solving the problem of being unable to simultaneously address single-point faults and batch fault identification.

[0075] For example, the final judgment result can be used to generate a structured work order. At the same time, the manually verified labels are fed back into the training dataset, and the model is retrained periodically (e.g., weekly) to achieve closed-loop optimization.

[0076] For example, the following two examples illustrate this: Example 1: Identification of Potential Hidden Dangers in Downlink Fiber Optic Cables The feature vector of the first user is: [1, 1, 1, 1, 4, 0.17, 8, -28.3]. The probability output by the machine learning model is 0.72 (greater than the preset threshold of 0.6). The system determines that the first and second preset conditions are met, but the third preset condition is not met, and therefore identifies it as a "potential fiber optic vulnerability".

[0077] Example 2: Identification of Potential Hazards in Optical Splitters

[0078] The feature vector of the second user is: [1, 1, 1, 1, 4.2, 0.67, 45, -25.1]. The model output probability is 0.85 (greater than the preset threshold of 0.6). After judging that the first and third preset conditions are met, it is finally determined to be a potential problem with the secondary optical splitter.

[0079] It is understood that the various method embodiments mentioned above in this disclosure can be combined with each other to form combined embodiments without violating the principle and logic. Due to space limitations, this disclosure will not elaborate further. Those skilled in the art will understand that in the above methods of specific implementation, the specific execution order of each step should be determined by its function and possible internal logic.

[0080] In addition, this disclosure also provides an intelligent identification system, electronic device, and computer-readable storage medium for fiber optic network vulnerability types. All of the above can be used to implement any of the intelligent identification methods for fiber optic network vulnerability types provided in this disclosure. The corresponding technical solutions and descriptions are described in the corresponding descriptions in the method section and will not be repeated here.

[0081] Figure 6 This is a block diagram of an intelligent identification system for fiber optic network vulnerability types, provided as an embodiment of the present disclosure.

[0082] Reference Figure 6 This disclosure provides an intelligent identification system for fiber optic network hazard types. The intelligent identification system 600 for fiber optic network hazard types includes a model acquisition module 601, a real-time data acquisition module 602, a probability determination module 603, and a hazard type determination module 604.

[0083] The model acquisition module 601 is configured to acquire historical data related to offline historical alarm events of the ONU. The classification model is trained based on this historical data to obtain a hazard assessment model.

[0084] The real-time data acquisition module 602 is configured to: monitor the status of the ONU in real time, and acquire real-time data related to the current ONU offline alarm event when an ONU offline alarm event occurs.

[0085] The probability determination module 603 is configured to: obtain the probability that the current ONU offline alarm event is a hidden danger based on real-time data related to the current ONU offline alarm event using a hidden danger judgment model.

[0086] The hazard type determination module 604 is configured to: determine whether the current ONU offline alarm event is a hazard based on the probability that it is a hazard and a preset threshold. If so, determine the hazard type of the current ONU offline alarm event according to preset rules.

[0087] Figure 7 This is a block diagram of an electronic device provided in an embodiment of the present disclosure.

[0088] Reference Figure 7 This disclosure provides an electronic device, which includes: at least one processor 701; at least one memory 702; and one or more I / O interfaces 703 connected between the processor 701 and the memory 702; wherein the memory 702 stores one or more computer programs that can be executed by the at least one processor 701, and the one or more computer programs are executed by the at least one processor 701 to enable the at least one processor 701 to execute the above-described intelligent identification method for fiber optic network vulnerability types.

[0089] This disclosure also provides a computer-readable storage medium storing a computer program thereon, wherein the computer program, when executed by a processor, implements the aforementioned intelligent identification method for fiber optic network vulnerability types. The computer-readable storage medium can be volatile or non-volatile.

[0090] This disclosure also provides a computer program product, including computer-readable code, or a non-volatile computer-readable storage medium carrying computer-readable code. When the computer-readable code is run in the processor of an electronic device, the processor in the electronic device executes the above-described intelligent identification method for fiber optic network vulnerability types.

[0091] Those skilled in the art will understand that all or some of the steps, systems, and functional modules / units within the methods disclosed above can be implemented as software, firmware, hardware, or suitable combinations thereof. In hardware implementations, the division between functional modules / units mentioned above does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed collaboratively by several physical components. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit (ASIC). Such software may be distributed on a computer-readable storage medium, which may include computer storage media (or non-transitory media) and communication media (or transient media).

[0092] As is known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information, such as computer-readable program instructions, data structures, program modules, or other data. Computer storage media includes, but is not limited to, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), static random access memory (SRAM), flash memory or other memory technologies, portable compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage and systems, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, it is known to those skilled in the art that communication media typically contain computer-readable program instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0093] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0094] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.

[0095] The computer program product described herein can be implemented specifically through hardware, software, or a combination thereof. In one alternative embodiment, the computer program product is specifically embodied in a computer storage medium; in another alternative embodiment, the computer program product is specifically embodied in a software product, such as a software development kit (SDK), etc.

[0096] Various aspects of this disclosure are described herein with reference to flowchart illustrations and / or block diagrams of methods, systems (systems), and computer program products according to embodiments of this disclosure. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0097] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing and systems to produce a machine such that, when executed by the processor of the computer or other programmable data processing and systems, they create a system that implements the functions / actions specified in one or more blocks of a flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing and systems, and / or other devices to function in a particular manner. Thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of a flowchart and / or block diagram.

[0098] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing and system, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing and system, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing and system, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0099] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those shown in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0100] Example embodiments have been disclosed herein, and while specific terminology has been used, it is for illustrative purposes only and should be construed as such, and is not intended to be limiting. In some instances, it will be apparent to those skilled in the art that features, characteristics, and / or elements described in connection with particular embodiments may be used alone, or in combination with features, characteristics, and / or elements described in connection with other embodiments, unless otherwise expressly indicated. Therefore, those skilled in the art will understand that various changes in form and detail may be made without departing from the scope of this disclosure as set forth by the appended claims.

Claims

1. A method for intelligent identification of potential risks in fiber optic networks, characterized in that, include: S1, obtain historical data related to historical alarm events that are offline with the ONU; The classification model is trained based on the historical data to obtain a hazard assessment model; S2 monitors the status of the ONU in real time and obtains real-time data related to the current ONU offline alarm event when an ONU offline alarm event occurs. S3, Based on the real-time data related to the current ONU offline alarm event, use the hidden danger judgment model to obtain the probability that the current ONU offline alarm event is a hidden danger; S4, determine whether the current ONU offline alarm event is a potential hazard based on the probability that the current ONU offline alarm event is a potential hazard and a preset threshold; If so, determine the potential problem type of the current ONU offline alarm event according to the preset rules.

2. The intelligent identification method for fiber optic network vulnerability types according to claim 1, characterized in that, S1 includes: S11, collect historical alarm events of multiple ONU offline within a preset historical time period, and obtain historical data related to the historical alarm events of multiple ONU offline and the hidden danger types marked by operation and maintenance experts for each historical alarm event of ONU offline; The identified hazard types include: fiber optic cable hazard, optical splitter hazard, or no hazard. Historical data related to multiple ONU offline alarm events includes: whether the current alarm event is an ONU offline alarm, whether the current alarm event occurred during working hours, whether the user involved in the current alarm event resides in an urban or rural residential area, whether the ONU model involved in the current alarm event is on an exclusion list, the number of times the ONU involved in the current alarm event went offline that day, the aggregation degree of the primary / secondary splitter to which the ONU involved in the current alarm event belongs, the duration of a single offline event in the current alarm event, and the optical attenuation value most recently reported by the ONU related to the current alarm event. The aggregation degree of the primary / secondary splitter to which the ONU involved in the current alarm event belongs is the maximum of the percentage of users whose primary splitter to which the ONU involved in the current alarm event has experienced more than 3 offline events and the percentage of users whose secondary splitter to which the ONU involved in the current alarm event has experienced more than 3 offline events. S12, construct historical feature vectors based on historical data related to the historical alarm events of multiple ONUs offline; combine each historical feature vector with the hidden danger type marked by the corresponding operation and maintenance expert for each ONU offline historical alarm event to obtain a feature dataset; S13, Train the classification model based on the feature dataset to obtain the hidden danger judgment model.

3. The intelligent identification method for fiber optic network vulnerability types according to claim 2, characterized in that, Historical feature vectors consist of feature values ​​in eight dimensions; among them, In the first dimension, if the alarm event is an ONU offline alarm, the feature value of the first dimension is 1; otherwise, the feature value of the first dimension is 0. In the second dimension, if the alarm event occurs during working hours, the feature value of the second dimension is 1; otherwise, the feature value of the second dimension is 0. In the third dimension, if the user involved in this alarm event belongs to an urban or rural residential community, the feature value of the third dimension is 1; otherwise, the feature value of the third dimension is 0. In the fourth dimension, if the ONU model involved in this alarm event is not in the exclusion list, the feature value of the fourth dimension is 1; otherwise, the feature value of the fourth dimension is 0. In the fifth dimension, the feature value of the fifth dimension is equal to the number of times the ONU involved in this alarm event went offline on that day; In the sixth dimension, the eigenvalue of the sixth dimension is equal to the aggregation degree of the first-level / second-level splitter to which the ONU involved in this alarm event belongs; In the seventh dimension, the feature value of the seventh dimension is equal to the single offline time in this alarm event. In the eighth dimension, the feature value of the eighth dimension is equal to the optical attenuation value most recently reported by the ONU related to this alarm event.

4. The intelligent identification method for fiber optic network vulnerability types according to claim 2, characterized in that, S3 include: S31, extract features from the real-time data related to the current ONU offline alarm event to obtain the feature vector of the real-time data related to the current ONU offline alarm event; S32, input the feature vector of the real-time data related to the current ONU offline alarm event into the hidden danger judgment model; the output of the hidden danger judgment model is the probability that the current ONU offline alarm event is a hidden danger.

5. The intelligent identification method for fiber optic network vulnerability types according to claim 2, characterized in that, S4 includes: S41, compare the probability that the current ONU offline alarm event is a potential hazard with the corresponding preset threshold to determine whether the current ONU offline alarm event is a potential hazard; S42, determine the potential risk type of the current ONU offline alarm event based on whether the current ONU offline alarm event is a potential risk and the preset rules.

6. The intelligent identification method for fiber optic network vulnerability types according to claim 5, characterized in that, S42 includes: S421, determine whether the features in the feature vector of the real-time data related to the current ONU offline alarm event meet the first preset condition; if yes, execute S422, otherwise execute S424; the first preset condition includes: the alarm event is an ONU offline alarm, the alarm event occurs during working hours, the user involved in the alarm event belongs to an urban or rural residential community, the ONU model involved in the alarm event is not in the exclusion list, and the ONU involved in the alarm event has more than the preset offline number of times on the same day; S422, determine whether the features in the feature vector of the real-time data related to the current ONU offline alarm event meet the second preset condition; if yes, determine that the hidden danger type of the current ONU offline alarm event is the fiber optic hidden danger; otherwise, execute S423; the second preset condition includes: the optical attenuation value most recently reported by the ONU related to this alarm event is less than the preset optical attenuation value; S423, determine whether the features in the feature vector of the real-time data related to the current ONU offline alarm event meet the third preset condition; if yes, determine that the hidden danger type of the current ONU offline alarm event is an optical splitter hidden danger; otherwise, execute S424; the third preset condition includes: the aggregation degree of the first-level / second-level optical splitter to which the ONU involved in this alarm event belongs is greater than the preset aggregation degree, and the total number of users connected to the ONU involved in this alarm event is greater than or equal to the preset number of users, and the single offline time in this alarm event is within the preset offline time period; S424 generates a work order and pushes it to the administrator for manual verification.

7. An intelligent identification system for types of hidden dangers in fiber optic networks, characterized in that, include: The model acquisition module is configured to: acquire historical data related to offline historical alarm events of the ONU; train the classification model based on the historical data to obtain a hidden danger judgment model; The real-time data acquisition module is configured to: monitor the status of the ONU in real time, and acquire real-time data related to the current ONU offline alarm event when an ONU offline alarm event occurs; The probability determination module is configured to: based on the real-time data related to the current ONU offline alarm event, use the hidden danger judgment model to obtain the probability that the current ONU offline alarm event is a hidden danger; The hazard type determination module is configured to: determine whether the current ONU offline alarm event is a hazard based on the probability that the current ONU offline alarm event is a hazard and a preset threshold; if so, determine the hazard type of the current ONU offline alarm event according to preset rules.

8. An electronic device, characterized in that, include: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores one or more computer programs that can be executed by the at least one processor, and the one or more computer programs are executed by the at least one processor to enable the at least one processor to perform the intelligent identification method for fiber optic network vulnerability types as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the intelligent identification method for fiber optic network vulnerability types as described in any one of claims 1-6.

10. A computer program product, characterized in that, Includes computer-readable code, or a non-volatile computer-readable storage medium carrying computer-readable code, wherein when the computer-readable code is run in a processor of an electronic device, the processor in the electronic device executes an intelligent identification method for fiber optic network vulnerability types as described in any one of claims 1-6.

Images