Fault early warning and quick response handling method based on network equipment operation and maintenance data

Abstract

The invention discloses a fault early warning and quick response processing method based on network equipment operation and maintenance data, and relates to the technical field of network equipment operation and maintenance. According to the method, network equipment hardware, software and link state data and key component energy consumption data are collected in a multi-dimensional mode, a correlation model and a rule base are constructed, quantitative grading and correlation analysis are combined, fault hidden dangers and correlation components are accurately recognized, fault early warning is achieved, an adaptive response disposal scheme is generated based on the hidden danger level, and the fault handling efficiency is improved. The execution safety is ensured through instruction verification, the fault handling efficiency is greatly improved, and the operation and maintenance cost and the fault influence are reduced; meanwhile, on the premise of guaranteeing safety, the equipment operation state is optimized, the energy consumption of key components is regulated and controlled to the optimal interval, and the energy utilization rate and the equipment operation stability are improved; according to the invention, fault early warning precision, response processing high efficiency and operation state optimization are realized, and the intelligent level and comprehensive benefits of network equipment operation and maintenance are significantly improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of network equipment operation and maintenance, and particularly relates to a fault early warning and rapid response handling method based on network equipment operation and maintenance data. BACKGROUND

[0002] Existing operation and maintenance relies on manual inspection or single-dimensional data monitoring, which is difficult to comprehensively capture the equipment running state and energy consumption changes, resulting in untimely fault early warning and inaccurate hidden danger positioning. For example, the Chinese patent with publication number CN118200118A discloses a substation communication network equipment monitoring and fault early warning method and system, which comprises: data collection and preprocessing, collecting substation communication network equipment data; extracting useful features from the preprocessed data to support subsequent model training, feature extraction is realized through a denoising autoencoder; a model for predicting equipment failure is constructed based on an LSTM network and trained; the trained model is used to monitor the network state in real time, and abnormal or potential problems are identified; once a problem is detected, the early warning mechanism will start, automatically sending warning information to the operation and maintenance personnel through the predetermined communication channel, so that they can respond and solve the detected problem in time, ensuring the stable and safe operation of the substation communication network. This technical solution combines the advantages of LSTM network and denoising autoencoder to realize efficient and accurate monitoring and early warning.

[0003] The above patent application improves the accuracy of substation communication network equipment monitoring and early warning by means of LSTM network and denoising autoencoder, but still has the following problems:

[0004] 1. The existing technology has a single data monitoring dimension, only focusing on equipment communication related data collection and analysis, without correlating and fusing equipment key component energy consumption data and running state data, and unable to use energy consumption abnormal features to reverse prove equipment fault hidden dangers, resulting in incomplete fault cause judgment;

[0005] 2. The existing technology does not optimize data collection strategies for different operating conditions, lacks parameter calibration and rule base construction that are self-adaptive to operating conditions, and is difficult to adapt to monitoring needs in different scenarios such as peak load and normal operation, which may result in inaccurate data collection in high load conditions and resource waste in low load conditions;

[0006] 3. The existing technology only realizes the function of fault early warning, lacks adaptive disposal schemes for different fault types, and does not consider the optimization of equipment operation after early warning, cannot balance fault rapid disposal and energy utilization efficiency improvement, and when facing multi-task concurrent scenarios, may delay key fault disposal and cause unreasonable resource allocation. SUMMARY

[0007] The purpose of the present application is to provide a network equipment operation and maintenance data-based fault early warning and rapid response handling method, which collects equipment state and energy consumption data in multiple dimensions, combines correlation analysis and quantitative classification, predicts fault risks in advance, generates targeted early warning prompts and response plans, optimizes key component operating conditions for scenarios without hidden dangers or low-level hidden dangers on the basis of ensuring safe operation of equipment, improves energy utilization and equipment operating efficiency, realizes the synergy of safety and energy saving, and solves the problems raised in the above background technology.

[0008] To achieve the above purpose, the present application provides the following technical solutions:

[0009] The network equipment operation and maintenance data-based fault early warning and rapid response handling method comprises the following steps:

[0010] Visual and energy consumption perception: real-time capture of relevant data of network equipment hardware state, software running state and network topology link state; synchronous acquisition of target energy consumption data of network equipment key components, establishment of the correlation between equipment state data and network equipment key component target energy consumption data, and generation of perception data set;

[0011] Fault correlation analysis: obtain the perception data set, extract the target feature parameters in the perception data set and perform quantitative classification, and at the same time, perform trend analysis on the target energy consumption data to extract abnormal energy consumption features; analyze the correlation degree of the target feature parameters and the abnormal energy consumption features, determine the fault hidden danger type and the target key components or link nodes causing energy consumption abnormalities and fault hidden dangers, and generate a fault correlation analysis report;

[0012] Early warning and response decision generation: based on the fault correlation analysis report, in combination with the correlation degree of the target feature parameters and the abnormal energy consumption features corresponding to the perception data set, judge the operation optimization space of the target key components, generate early warning prompts and response handling plans adapted to different fault hidden danger types, and update the network equipment operating parameter benchmark in synchronization;

[0013] Early warning response implementation and optimization: obtain the early warning prompts and response handling plans, the updated operating parameter benchmark, perform targeted handling and regulation on the target key components or link nodes, and synchronously obtain regulation feedback data, dynamically correct the correlation of the perception data set based on the regulation feedback data, and calibrate the handling and regulation accuracy.

[0014] Further, the visual and energy consumption perception further comprises presetting perception nodes based on the operating conditions of the network equipment, laying data acquisition equipment on the key components of the network equipment core processor, power module, heat dissipation system, network interface card and network topology key link nodes, and laying energy consumption acquisition units on the key components.

[0015] Further, the specific process of generating the perception data set comprises:

[0016] Based on the load characteristics and operational requirements of network devices under different operating conditions, dynamic parameter calibration is performed on the data acquisition devices at preset sensing nodes to obtain data acquisition adaptation parameters.

[0017] Based on the energy consumption characteristics of each target key component, the range and accuracy of the pre-deployed energy consumption acquisition units are adapted and calibrated, and differentiated data acquisition frequencies and unified timestamps are set to obtain accurate energy consumption acquisition parameters.

[0018] The data acquisition device captures various equipment status data in real time according to the adapted parameters, extracts the initial status characteristics, and the energy consumption acquisition unit collects the energy consumption data of each target key component in real time according to the precise parameters.

[0019] By combining the energy consumption baseline range corresponding to the operating conditions, abnormal data is removed and effective energy consumption parameters are retained to form a preprocessed data set of state characteristics and energy consumption parameters;

[0020] Based on the preprocessed dataset and combined with a unified timestamp, a one-to-one correspondence between status data and energy consumption data is established. At the same time, real-time operating condition tags of network devices are obtained, and a correlation model between status data, energy consumption data and real-time operating condition of network devices is established.

[0021] Based on the correlation model, the inherent correlation patterns between state data and energy consumption data under different operating conditions are mined to generate an adaptive correlation rule base for operating conditions, thus forming structured correlation data.

[0022] Based on structured associated data and an adaptive association rule base for operating conditions, the validity of the data and the rationality of the association are verified, and a perception dataset is generated based on the verification results.

[0023] Furthermore, the specific process of quantitative grading includes:

[0024] Based on the adaptive association rule base of working conditions, target feature parameters such as excessive CPU temperature, abnormal memory usage, number of bad sectors on hard disk, number of port connection interruptions, link packet loss rate, and protocol error frequency are extracted from the perception dataset under the corresponding working conditions.

[0025] Based on network equipment security operation specifications and historical fault data, the security weights corresponding to each target characteristic parameter are determined, and a quantitative hierarchical evaluation matrix is ​​constructed.

[0026] Cluster analysis is performed on the target feature parameters based on the quantitative grading evaluation matrix to obtain the grade interval corresponding to each target feature parameter and determine the safety level of each target feature parameter.

[0027] The quantitative grading results are coupled with the abnormal energy consumption characteristics in the perception dataset for analysis. The correlation coefficient between the target feature parameters of each security level and the energy consumption anomaly type is output, and a mapping relationship table between the grading results and energy consumption anomalies is established.

[0028] Historical operating data is acquired, and the accuracy of the classification results and the rationality of the correlation coefficients are verified based on the historical operating data. Based on the verification results, the probability of failure corresponding to the target characteristic parameters of different safety levels is determined in the mapping relationship table between the classification results and energy consumption anomalies.

[0029] Furthermore, the correlation between target characteristic parameters and abnormal energy consumption characteristics is analyzed, specifically including:

[0030] Obtain the mapping relationship table between the classification results and energy consumption anomalies, as well as the probability of failure corresponding to each safety level. Combine the working condition adaptive association rule library to extract the safety level and corresponding energy consumption anomaly type of the target feature parameters under the current working condition.

[0031] Using the safety level of the target feature parameters as a reference sequence and the corresponding energy consumption anomaly type and failure probability as a comparison sequence, the correlation value between the target feature parameters of each safety level and different energy consumption anomaly types is calculated.

[0032] Based on the network equipment security operation specifications, a correlation threshold is set, the calculated correlation value is compared with the correlation threshold, the correlation level is divided, and the priority of the corresponding energy consumption anomalies and faults caused by the target characteristic parameters of different levels is clarified.

[0033] By combining the current operating condition labels and the operating condition adaptive association rule library, the association level is adjusted to adapt to the operating condition, and the corresponding relationship between the corrected target feature parameters, energy consumption anomaly types and association levels is output.

[0034] Furthermore, using the safety level of the target characteristic parameters as a reference sequence and the corresponding energy consumption anomaly type and failure probability as a comparison sequence, the correlation value between the target characteristic parameters of each safety level and different energy consumption anomaly types is calculated, including:

[0035] Determine the probability of fault occurrence for each target feature parameter corresponding to the energy consumption anomaly type based on the reference sequence and comparison sequence;

[0036] The probability of fault occurrence for the energy consumption anomaly type corresponding to the target feature parameter is normalized to obtain the normalized fault occurrence probability.

[0037] The probability confidence function is used to process the normalized fault occurrence probability to obtain the fault probability coefficients under the energy consumption anomaly type corresponding to the target feature parameters;

[0038] The actual occurrence rate of energy consumption anomaly types corresponding to the target feature parameters is retrieved from historical fault data, and an observation sequence is generated using the actual occurrence rate of energy consumption anomaly types corresponding to the target feature parameters.

[0039] The similarity between the observed sequence and the comparison sequence corresponding to the target feature parameters is processed to obtain the similarity value between the observed sequence and the comparison sequence corresponding to the target feature parameters.

[0040] The correlation between the target feature parameters of each safety level and different energy consumption anomaly types is calculated by combining the similarity values ​​between the observation sequences and comparison sequences corresponding to the target feature parameters with the failure probability coefficients under the energy consumption anomaly types corresponding to the target feature parameters.

[0041] Furthermore, the observed sequences and comparison sequences corresponding to the target feature parameters are subjected to similarity processing to obtain the similarity values ​​between the observed sequences and comparison sequences corresponding to the target feature parameters, including:

[0042] Retrieve the number of energy consumption anomaly types contained in the observation and comparison sequences corresponding to the target feature parameters;

[0043] Retrieve the probability of failure and the actual percentage of occurrence for each type of energy consumption anomaly;

[0044] The probability of failure and the actual occurrence rate corresponding to each type of energy consumption anomaly are compared to obtain the absolute difference between the probability of failure and the actual occurrence rate.

[0045] The similarity value between the observation sequence and the comparison sequence corresponding to the target feature parameter is obtained by combining the absolute difference between the failure probability and the actual occurrence rate of each type of energy consumption anomaly contained in the observation sequence and the comparison sequence corresponding to the target feature parameter with the corresponding values ​​of the failure probability and the actual occurrence rate of each type of energy consumption anomaly.

[0046] Furthermore, the fault correlation analysis report includes: basic information, quantitative grading results, correlation analysis conclusions, fault hazard type, hazard level and corresponding target key component or link node information, and auxiliary decision-making data.

[0047] Furthermore, the early warning and response decision generation also includes:

[0048] The decision-making criteria prioritize safety over energy optimization, and the basic fault handling control instructions are generated by combining the hazard level information in the fault correlation analysis report.

[0049] Once scenarios with no potential faults or low-level potential faults are identified, operation optimization control commands are generated based on the correspondence between target characteristic parameters, energy consumption anomaly types, and correlation levels, thereby adjusting the operating status of each target key component to the optimal range of the corresponding operating conditions.

[0050] Establish instruction compatibility verification rules, cross-verify basic fault handling control instructions and operation optimization control instructions, eliminate conflicting control instructions, perform safety adaptation correction on boundary parameters, and finally generate an executable composite control instruction set.

[0051] Furthermore, the implementation and optimization of early warning response also include:

[0052] Obtain the network device operation and maintenance task description information corresponding to the response and handling plan, decompose the task description information into parameters, and extract the task target information, task type information, and task operation information.

[0053] Based on task objective information, combined with task type information and task operation information, a load coefficient calculation model is constructed to quantitatively evaluate the task load coefficient of each network device operation and maintenance task.

[0054] Based on the task load coefficient threshold classification standard, all network device operation and maintenance tasks are divided into light-load tasks and heavy-load tasks, and phase task indicators and overall task indicators are set respectively.

[0055] Based on the correlation between phase task indicators and overall task indicators and the expected change range during the operation process, a state transition matrix is ​​constructed for each segmented network device operation and maintenance task.

[0056] Based on the state transition matrix, the standard deduction task process parameters for each partitioned network device operation and maintenance task are determined, and core monitoring indicators are selected.

[0057] Match the corresponding handling and control response sub-tasks for each monitoring indicator, clarify the scheduling resources required for each response sub-task, and determine whether there are any synchronization domain tasks that need to be executed synchronously through resource occupation conflict detection;

[0058] If there are synchronization domain task items, retrieve the network device control program corresponding to each partitioned network device operation and maintenance task, and extract the spatial interleaving features of the synchronization domain task items in terms of time dimension and execution logic.

[0059] Based on the resource contention intensity and execution priority ranking of synchronous domain task items with spatial interleaved information characteristics, and combined with the resource constraints of network device operation, the resource requirements and priority weights for handling and regulation of each network device operation and maintenance task are determined.

[0060] Based on the urgency of loading space requirements and the priority weight of handling and control, the network equipment operation and maintenance tasks in the light-load operation task type and the heavy-load operation task type are sorted.

[0061] The execution order of the handling and control is determined based on the sorting results. Precise handling and control are carried out sequentially for each network device operation and maintenance task, and control feedback data is collected simultaneously during the execution of each task.

[0062] Compared with the prior art, the beneficial effects of the present invention are:

[0063] This invention collects multi-dimensional data on network device hardware, software, and link status, as well as energy consumption data of key components, to construct a correlation model and rule base. Combined with quantitative grading and correlation analysis, it accurately identifies potential faults and related components, enabling early fault warnings. Based on the fault level, it generates appropriate response and handling solutions, and ensures execution security through instruction verification, significantly improving fault handling efficiency and reducing maintenance costs and fault impact. Simultaneously, while ensuring safety, it optimizes equipment operating status, regulates the energy consumption of key components to the optimal range, and improves energy utilization and equipment operational stability. This achieves precise fault warnings, efficient response and handling, and optimized operating status, significantly improving the intelligence level and overall benefits of network device operation and maintenance. Attached Figure Description

[0064] Figure 1 This is a flowchart of the fault early warning and rapid response handling method based on network device operation and maintenance data of the present invention. Detailed Implementation

[0065] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0066] To address the shortcomings of existing equipment monitoring and early warning technologies, such as limited data monitoring dimensions, lack of optimized data acquisition strategies based on operating conditions, and the inability to provide fault warnings without adaptable handling solutions and operational optimization mechanisms, which lead to incomplete fault cause identification, poor adaptability to different scenarios, and delays in handling critical faults under multi-task concurrency, please refer to [link to relevant documentation]. Figure 1 This embodiment provides the following technical solution:

[0067] The fault early warning and rapid response handling method based on network equipment operation and maintenance data includes the following steps:

[0068] Visual and Energy Consumption Perception: Based on the operating conditions of network devices, such as peak load, normal operation, and low load, preset perception nodes are established. Data acquisition devices are deployed at key components of network devices, such as core processors, power modules, cooling systems, and network interface cards, as well as key link nodes in the network topology. This allows for real-time capture of network device hardware status, such as CPU temperature, memory usage, hard drive read / write speed, and port connection stability; software operating status, such as process resource usage, protocol operation status, and log error information; and network topology link status, such as link bandwidth utilization, data transmission latency, and packet loss rate. Simultaneously, energy consumption acquisition units are deployed at key components to collect target energy consumption data for these components, including real-time power, voltage, current, and energy loss data. A correlation is established between device status data and target energy consumption data of key network device components, generating a perception dataset containing status characteristics and energy consumption parameters, providing a multi-dimensional data source for subsequent analysis.

[0069] Fault Correlation Analysis: Employing feature point detection and contour extraction algorithms, this study extracts target feature parameters from the perception dataset, including excessive CPU temperature, abnormal memory usage, number of bad sectors on the hard drive, number of port connection interruptions, link packet loss rate, and protocol error frequency. These parameters are then quantified and graded. Simultaneously, trend analysis is performed on the target energy consumption data to extract abnormal energy consumption features such as sudden energy consumption changes, sustained high energy consumption, and energy consumption mismatch with load. A state feature-energy consumption parameter mapping model is constructed based on historical operating data. This model analyzes the correlation between target feature parameters and abnormal energy consumption features. For example, sustained excessive CPU temperature corresponds to abnormally high power module energy consumption, and a surge in link packet loss rate corresponds to fluctuating network interface card energy consumption. This identifies the type of potential fault, such as hardware failure, software failure, or link failure, as well as the target key components or link nodes that cause abnormal energy consumption and potential faults. A fault correlation analysis report is generated, including the fault level, abnormal energy consumption threshold, and information on associated components / links.

[0070] Early warning and response decision generation: Based on the fault correlation analysis report, combined with the correlation degree between the target feature parameters and abnormal energy consumption characteristics corresponding to the perception dataset, the operation optimization space of the target key components is judged. The network equipment safety operation specifications and operation optimization objectives are integrated to generate early warning prompts adapted to different fault hazard types, such as audible and visual warnings, system pop-up warnings, and SMS notification warnings, as well as response and handling solutions, such as component restart, parameter adjustment, link switching, and shutdown maintenance. The network equipment operation parameter benchmark is updated simultaneously to achieve collaborative optimization of the dual objectives of fault handling and operation optimization.

[0071] Early warning response implementation and optimization: Obtain early warning prompts and response plans, updated operating parameter benchmarks, and carry out targeted control and regulation of key components or link nodes that cause abnormal energy consumption and potential faults. Simultaneously, obtain control feedback data, such as component operating status data, energy consumption data, and network transmission quality data. Based on the control feedback data, dynamically correct the correlation of the sensing dataset, calibrate the control and regulation accuracy, and ensure the safe, efficient, and low-power operation of network equipment.

[0072] In this embodiment, multi-dimensional data collection and correlation construction breaks through the limitations of a single monitoring dimension, comprehensively covering the operating status of network devices and the energy consumption of core components, providing three-dimensional and comprehensive data support for fault early warning and operation optimization. Professional algorithms are used to deeply mine the perceived data, accurately extracting device status characteristic parameters and abnormal energy consumption characteristics. A mapping model constructed from historical data clarifies the intrinsic relationship between the two, enabling precise location of fault hazard types, levels, and associated components / links, improving the timeliness and reliability of fault early warning. Prioritizing fault handling is the core principle, integrating network device safety operation specifications and operation optimization goals. Customized and adapted early warning prompts and response handling schemes are developed for different hazard scenarios, and operating parameter benchmarks are updated synchronously to avoid operational imbalances caused by a single goal, achieving synergistic advancement of safety assurance and operation optimization. Through task layering, resource scheduling optimization, and dynamic feedback calibration, the targeted, effective, and continuous nature of handling and control is ensured. While ensuring the safe and stable operation of network devices, operational efficiency and energy utilization are improved, significantly reducing fault losses and maintenance costs, and enhancing the overall benefits and sustainability of network device operation and maintenance.

[0073] In this embodiment, the specific process of generating the perception dataset includes:

[0074] Based on the load characteristics and operational requirements of network devices under different operating conditions, dynamic parameter calibration is performed on the data acquisition devices at the preset sensing nodes to adjust the data acquisition frequency, acquisition accuracy threshold, etc., to ensure coverage of the core operating status of the devices and key link indicators, eliminate monitoring blind spots caused by operating condition switching, and obtain data acquisition adaptation parameters.

[0075] Based on the energy consumption characteristics of each target key component, the range and accuracy of the pre-deployed energy consumption acquisition units are adapted and calibrated. Differentiated data acquisition frequencies and unified timestamps are set to ensure the time synchronization and acquisition accuracy of status data and energy consumption data, and to obtain accurate energy consumption acquisition parameters.

[0076] The data acquisition device captures various equipment status data in real time according to the adaptive parameters. An adaptive filtering algorithm is used to remove environmental interference and instantaneous equipment fluctuation interference. After data preprocessing, the initial state characteristics are initially extracted. The energy consumption data of each target key component is collected synchronously according to the precise parameters through the energy consumption acquisition unit.

[0077] Based on the energy consumption baseline range corresponding to the operating conditions, statistical analysis methods are used to remove abnormal data, such as sudden data caused by data acquisition equipment failure, while retaining effective energy consumption parameters to form a preprocessed data set of state characteristics and energy consumption parameters.

[0078] Based on the preprocessed dataset and combined with a unified timestamp, a one-to-one correspondence between status data and energy consumption data is established. At the same time, real-time operating condition tags of network devices are obtained, and a correlation model between status data, energy consumption data and real-time operating condition of network devices is established.

[0079] Based on the association model, the inherent correlation between state data and energy consumption data under different operating conditions is explored. For example, the correlation between high CPU utilization and increased power module energy consumption under high load conditions is explored. An adaptive association rule library for operating conditions is generated to achieve accurate mapping between target state features and corresponding component energy consumption parameters. After labeling data attributes, structured association data is formed.

[0080] Based on structured associated data and an adaptive association rule base for working conditions, we perform data validity and association rationality checks to identify problems such as missing data and misaligned associations.

[0081] The missing data is intelligently completed using a working condition inference algorithm, and the incorrectly associated data is corrected based on a rule base. Based on the verification results, a perception dataset containing complete state features, effective energy consumption parameters, and accurate working condition coupling relationships is generated.

[0082] In this embodiment, by dynamically calibrating the parameters of the data acquisition equipment, the core monitoring components and link nodes are accurately covered, effectively avoiding monitoring loopholes caused by operating condition switching and ensuring the comprehensiveness and relevance of status data acquisition. Differentiated acquisition frequencies and unified timestamps are set to achieve precise temporal synchronization of status data and energy consumption data, ensuring that the acquisition accuracy of the two types of data is mutually compatible. Through data preprocessing and verification, effective features and parameters are extracted to improve data quality and reduce interference from invalid information in subsequent analysis processes. A close correlation is established between status data, energy consumption data, and operating conditions, uncovering the inherent correlation patterns in the data and forming a rule base, achieving precise mapping between features and parameters, transforming the data from a scattered state into a structured relational form, and enhancing the application value of the data.

[0083] In this embodiment, the specific process of quantitative grading includes:

[0084] Based on the adaptive association rule base of working conditions, target feature parameters such as excessive CPU temperature, abnormal memory usage, number of bad sectors on hard disk, number of port connection interruptions, link packet loss rate, and protocol error frequency are extracted from the perception dataset under the corresponding working conditions.

[0085] Based on network equipment security operation specifications and historical fault data, the security weights corresponding to each target feature parameter are determined. For example, the weight of excessive CPU temperature is higher than the weight of slight abnormal memory usage. A quantitative grading evaluation matrix is ​​constructed. The row vectors of the quantitative grading evaluation matrix correspond to the target feature parameters, and the column vectors correspond to the three levels of safety, criticality, and danger.

[0086] Cluster analysis is performed on the target feature parameters based on the quantitative grading evaluation matrix to obtain the grade interval corresponding to each target feature parameter, determine the safety level of each target feature parameter, and calibrate the clustering results through the deviation correction algorithm to reduce the grading error caused by environmental interference.

[0087] For discrete target characteristic parameters such as the number of port connection interruptions and the number of bad sectors on the hard disk, the counting statistics method is used to quantify the grading results and label the specific values ​​and grade classifications. For continuous target characteristic parameters such as CPU temperature exceeding the standard and link packet loss rate, the interval mapping method is used to quantify the grading results and clarify the grade interval boundaries corresponding to the parameter values.

[0088] The quantitative grading results are coupled with the abnormal energy consumption characteristics in the perception dataset for analysis. The correlation coefficient between the target feature parameters of each security level and the energy consumption anomaly type is output, and a mapping relationship table between the grading results and energy consumption anomalies is established.

[0089] Historical operational data is acquired, and the accuracy of the classification results and the rationality of the correlation coefficients are verified based on the historical operational data. Based on the verification results, the probability of failure corresponding to the target characteristic parameters of different safety levels is calculated and determined in the mapping relationship table between the classification results and energy consumption anomalies through probability and statistical algorithms, so as to provide a classification basis for subsequent fault correlation analysis.

[0090] In this embodiment, the correlation between target characteristic parameters and abnormal energy consumption characteristics is analyzed, specifically including:

[0091] Obtain the mapping relationship table between the classification results and energy consumption anomalies, as well as the probability of failure corresponding to each safety level. Combine the working condition adaptive association rule library to extract the safety level and corresponding energy consumption anomaly type of the target feature parameters under the current working condition.

[0092] The grey relational algorithm is used to calculate the correlation between the target feature parameters of each safety level and different energy consumption anomaly types, with the safety level of the target feature parameters as the reference sequence and the corresponding energy consumption anomaly type and fault occurrence probability as the comparison sequence.

[0093] Based on the network equipment security operation specifications, a correlation threshold is set. The calculated correlation value is compared with the correlation threshold to divide the correlation into three levels: strong correlation, medium correlation, and weak correlation. The priority of the corresponding energy consumption anomalies and faults caused by the target characteristic parameters of different levels is clarified.

[0094] Combining the current operating condition labels and the operating condition adaptive association rule library, the association level is corrected to adapt to the operating condition, and false associations caused by operating condition interference are eliminated. For example, the brief energy consumption anomaly under instantaneous peak load has no substantial correlation with equipment failure. The corrected target feature parameters, energy consumption anomaly type and association level correspondence are output.

[0095] In this embodiment, using the security level of the target feature parameters as a reference sequence and the corresponding energy consumption anomaly type and failure probability as a comparison sequence, the correlation value between the target feature parameters of each security level and different energy consumption anomaly types is calculated, including:

[0096] Determine the probability of fault occurrence for each target feature parameter corresponding to the energy consumption anomaly type based on the reference sequence and comparison sequence;

[0097] The probability of fault occurrence for the energy consumption anomaly type corresponding to the target feature parameter is normalized to obtain the normalized fault occurrence probability.

[0098] The probability confidence function is used to process the normalized fault occurrence probability to obtain the fault probability coefficients under the energy consumption anomaly type corresponding to the target feature parameters;

[0099] The fault probability coefficient corresponding to the energy consumption anomaly type for the target feature parameter is obtained by the following formula:

[0100] Ф(p ij )=1-e -k×p ij

[0101] Among them, Ф(p ij ) represents the fault probability coefficient corresponding to the combination of the i-th target feature parameter and the j-th energy consumption anomaly type. The fault probability coefficient is a fault confidence index obtained by nonlinear transformation of the normalized fault occurrence probability through a probability confidence function. It is used to quantify the credibility of the fault occurrence under the combination of the i-th target feature parameter and the j-th energy consumption anomaly type. The larger the value of the fault probability coefficient, the higher the credibility of the fault occurrence, and vice versa. Ф() represents the probability confidence function. k represents the adjustment coefficient, which ranges from 1.8 to 3.0. It is used to control the sensitivity of the conversion from fault occurrence probability to fault probability coefficient. The larger the adjustment coefficient, the more sensitive the conversion from fault occurrence probability to fault probability coefficient is to low-probability fault events. It is suitable for scenarios with extremely high safety requirements. The smaller the adjustment coefficient, the more conservative the response of the conversion from fault occurrence probability to fault probability coefficient is to the fault probability. It is suitable for scenarios with low false alarm tolerance. p ij Let represent the probability of failure under the combination of the i-th target feature parameter and the j-th energy consumption anomaly type; e represents the base of the exponential function;

[0102] The actual occurrence rate of energy consumption anomaly types corresponding to the target feature parameters is retrieved from historical fault data, and an observation sequence is generated using the actual occurrence rate of energy consumption anomaly types corresponding to the target feature parameters.

[0103] The similarity between the observed sequence and the comparison sequence corresponding to the target feature parameters is processed to obtain the similarity value between the observed sequence and the comparison sequence corresponding to the target feature parameters.

[0104] The correlation between the target feature parameters of each safety level and different energy consumption anomaly types is calculated by combining the similarity values ​​between the observation sequences and comparison sequences corresponding to the target feature parameters with the failure probability coefficients under the energy consumption anomaly types corresponding to the target feature parameters.

[0105] The correlation values ​​between the target feature parameters and different energy consumption anomaly types are obtained through the following formula:

[0106] R ij =[Ф(p ij )×S i ] / [Ф(p ij )×Y ij ] 1 / 2

[0107] Among them, R ij This represents the correlation value between the i-th target feature parameter and the j-th energy consumption anomaly type; Ф(p ij Y represents the failure probability coefficient corresponding to the combination of the i-th target feature parameter and the j-th energy consumption anomaly type; ij S represents the actual occurrence rate of the i-th target feature parameter and the j-th type of energy consumption anomaly; i This represents the similarity value between the observed sequence and the comparison sequence corresponding to the i-th target feature parameter.

[0108] In this embodiment, the above-mentioned technical solution achieves scientific quantification and confidence enhancement of fault probability. By normalizing the fault occurrence probability, the dimensional differences of probability data in different dimensions are eliminated, ensuring data comparability. Combining the normalized probability with a probability confidence function with adjustment coefficients allows for flexible adjustment of the response sensitivity of the probability coefficients, effectively enhancing the confidence of the fault probability and making the quantification results more consistent with the actual fault occurrence patterns, avoiding the one-sidedness of the original probability data. Simultaneously, it effectively improves the comprehensiveness and relevance of the correlation analysis. By introducing the actual occurrence rate of historical fault data to generate an observation sequence, and performing similarity processing between this sequence and the comparison sequence, the correlation calculation incorporates historical actual operating patterns, breaking the limitations of analysis solely relying on fault probability, and making the correlation analysis between feature parameters and energy consumption anomaly types more consistent with actual application scenarios. Meanwhile, the aforementioned technical solution can also achieve accurate comprehensive calculation of correlation values. By integrating fault probability coefficients, actual occurrence rates, and sequence similarity values ​​to calculate correlation, multiple indicators are weighted and mutually verified. This considers both the probabilistic characteristics of fault occurrence and historical actual occurrences and sequence matching, ensuring that the final correlation value accurately reflects the true degree of correlation between target characteristic parameters and energy consumption anomaly types. Furthermore, it provides precise basis for energy consumption anomaly diagnosis and safety management. The correlation values ​​between characteristic parameters of different safety levels and energy consumption anomaly types can clearly define the degree of influence of each parameter on different energy consumption anomalies, helping to quickly locate key characteristic parameters of energy consumption anomalies and improving the targeting and efficiency of energy consumption anomaly identification. Simultaneously, the aforementioned technical solution allows for adjustment of the probability confidence function coefficient according to actual operating conditions, enabling the system to adapt to the energy consumption monitoring needs of different scenarios. The quantitative analysis process and clear calculation logic ensure that the correlation analysis results are reproducible and verifiable, providing scientific and reliable quantitative support for subsequent energy consumption fault early warning and safety level optimization, thus comprehensively improving the intelligence and refinement of energy consumption anomaly management.

[0109] Specifically, the similarity between the observed and compared sequences corresponding to the target feature parameters is processed to obtain the similarity values ​​between them, including:

[0110] Retrieve the number of energy consumption anomaly types contained in the observation and comparison sequences corresponding to the target feature parameters;

[0111] Retrieve the probability of failure and the actual percentage of occurrence for each type of energy consumption anomaly;

[0112] The probability of failure and the actual occurrence rate corresponding to each type of energy consumption anomaly are compared to obtain the absolute difference between the probability of failure and the actual occurrence rate.

[0113] The similarity value between the observation sequence and the comparison sequence corresponding to the target feature parameter is obtained by combining the absolute difference between the failure probability and the actual occurrence rate of each type of energy consumption anomaly contained in the observation sequence and the comparison sequence corresponding to the target feature parameter with the corresponding values ​​of the failure probability and the actual occurrence rate of each type of energy consumption anomaly.

[0114] The similarity value between the observed sequence and the comparison sequence corresponding to the target feature parameter is obtained by the following formula:

[0115]

[0116] Among them, S i denoted by , J represents the similarity value between the observed sequence and the comparison sequence corresponding to the i-th target feature parameter. A higher similarity value indicates a higher overall similarity between the parameter and various anomaly types, suggesting that the parameter's behavior is generally consistent with the prediction, reflecting a high level of reliability in fault prediction. Conversely, a lower similarity value indicates a lower overall similarity, suggesting poor consistency between the parameter's behavior and the prediction, reflecting a lower level of reliability in fault prediction; L represents the number of energy consumption anomaly types contained in the observed sequence and the comparison sequence corresponding to the i-th target feature parameter; J ij M represents the smaller of the probability of failure and the actual proportion of failure corresponding to the j-th type of energy consumption anomaly contained in the observation sequence and comparison sequence of the i-th target feature parameter; ij C represents the larger of the probability of failure and the actual proportion of failure corresponding to the j-th type of energy consumption anomaly contained in the observation sequence and comparison sequence of the i-th target feature parameter; ij It represents the absolute difference between the probability of failure and the actual proportion of failures corresponding to the j-th type of energy consumption anomaly contained in the observation sequence and comparison sequence corresponding to the i-th target feature parameter.

[0117] In this embodiment, the above-mentioned technical solution effectively improves the comprehensiveness and relevance of similarity calculation. Based on the types of energy consumption anomalies, it covers all anomaly types to carry out similarity analysis, avoiding the one-sidedness of single-type analysis. At the same time, it focuses on two core data points: the probability of failure and the actual proportion of occurrence for each type, so that the similarity calculation closely follows the key characteristics of energy consumption anomalies and conforms to actual operating patterns.

[0118] Secondly, the accuracy and reliability of similarity values ​​are enhanced. By calculating the absolute difference between the probability of a fault occurrence and the actual proportion of occurrences, the deviation between the theoretical probability and historical data is quantified intuitively. This is then combined with the original values ​​of both types of data for comprehensive calculation, reflecting both the degree of data deviation and the influence of the original data's weights, thus avoiding distortion caused by judging similarity solely based on deviation values. Furthermore, this embodiment obtains accurate similarity values ​​through the above method, and these highly accurate similarity values ​​objectively reflect the degree of matching between the observed sequence (historical actual situation) and the comparison sequence (theoretical probability situation), providing a reliable basis for the integration of historical patterns and theoretical probabilities in correlation calculations, effectively improving the scientific rigor of overall energy consumption anomaly correlation analysis.

[0119] In this embodiment, the fault correlation analysis report includes:

[0120] Basic Information: Record the current network device operating condition label, the period and source of the sensing dataset collection, specify the data collection adaptation parameters, the energy consumption collection precision parameters and unified timestamp synchronization information, mark the operating condition adaptive association rule base version, and retain key records of data preprocessing and verification completion;

[0121] Quantitative grading results: The final quantitative grading results, corresponding safety levels and failure probabilities of each target feature parameter are included. The counting and statistical methods for discrete parameters and the interval mapping methods for continuous parameters are clarified. The grade interval boundaries, specific numerical labels and deviation correction instructions are attached.

[0122] The correlation analysis conclusions state the correlation values ​​between the target characteristic parameters of each safety level and the energy consumption anomaly type, the correlation level classification results and the correlation threshold, record the corresponding relationship between the parameters, energy consumption anomaly types and correlation levels after the working condition adaptation correction, and explain the basis for eliminating false correlations and the correction process.

[0123] Fault hazard type, hazard level and corresponding target key component or link node information: Determine the fault hazard type, such as hardware failure, software failure, link failure, and corresponding hazard level; mark the name, installation location and link node location and energy consumption characteristics of the target key component that causes energy consumption anomaly; record in detail the specific manifestations and occurrence sequence of abnormal characteristics such as energy consumption anomaly threshold, energy consumption mutation, and continuous high energy consumption.

[0124] Decision Support Data: This data summarizes the probability of failure and the conclusions of operating condition adaptation for each correlation level, and includes raw data and trend analysis curves of energy consumption fluctuations of key target components. This provides complete data support for the subsequent generation of early warning prompts and response plans, as well as for updating the benchmark operating parameters of network equipment.

[0125] In this embodiment, by accurately extracting target feature parameters and scientifically quantifying and classifying them, and combining historical data verification to ensure the accuracy and reliability of the classification results, a quantitative basis for fault hazard assessment is provided. Professional algorithms are used to analyze the correlation between target feature parameters and abnormal energy consumption characteristics, clarify the priority of fault occurrence, eliminate false correlation interference, and achieve accurate positioning of fault hazard type, level, and associated components / links, thereby improving the timeliness and reliability of fault early warning. The generated fault correlation analysis report contains multi-dimensional key information, providing comprehensive and accurate data support for subsequent decision-making and avoiding blind decision-making.

[0126] In this embodiment, the early warning and response decision generation also includes:

[0127] The decision criteria prioritize fault handling over operation optimization. Based on the hazard level information in the fault correlation analysis report, basic fault handling control instructions are generated, including faulty component isolation, abnormal link disconnection, and emergency shutdown.

[0128] Once scenarios with no or low-level potential hazards are identified, operation optimization control instructions are generated based on the correspondence between target characteristic parameters, energy consumption anomaly types, and associated levels. By adjusting parameters such as the core processor operating frequency, power module output power, and link bandwidth allocation strategy, the operating status of each target key component is controlled to the optimal range of the corresponding operating conditions while meeting network operating efficiency requirements. For example, the processor operating frequency is reduced and the energy consumption of the heat dissipation system is optimized under light load conditions.

[0129] Establish instruction compatibility verification rules, cross-verify basic fault handling control instructions and operation optimization control instructions, eliminate conflicting control instructions, and perform safety adaptation corrections on boundary parameters to ensure that the corrected instructions meet both the operation optimization objectives and the network device safety operation specifications, ultimately generating an executable composite control instruction set.

[0130] In this embodiment, targeted basic fault handling control instructions are generated based on the hazard level in the fault correlation analysis report. This mitigates security risks at the source, strengthens the security defense of network equipment operation, and ensures equipment stability and data security. In scenarios with no fault hazards or low-level hazards, the focus is on operational optimization needs. By dynamically adjusting key operating parameters, the energy consumption of each component is reasonably controlled while ensuring network operating efficiency, thereby improving energy utilization efficiency and reducing unnecessary energy consumption. By cross-validating and eliminating conflicting instructions and correcting boundary parameters, the collaborative adaptation of fault handling and operational optimization instructions is achieved. This avoids operational contradictions caused by a single objective, ensuring that the instructions comply with safety specifications and achieve operational optimization goals, thus improving the scientific nature and executability of decision-making instructions.

[0131] In this embodiment, the early warning response implementation and optimization also includes:

[0132] Obtain the network equipment operation and maintenance task description information corresponding to the response and handling plan, decompose the task description information into parameters, extract task target information such as fault repair, performance optimization, and energy consumption reduction, task type information such as emergency handling, routine maintenance, and upgrade optimization, and task operation information such as component replacement, parameter configuration, and program update, and clarify the core requirements and execution boundary conditions of network equipment operation and maintenance.

[0133] Based on parameters such as network load requirements and fault impact range in the task objective information, combined with task type information and task operation information, a load coefficient calculation model is constructed to quantitatively evaluate the task load coefficient of each network device operation and maintenance task, reflecting the comprehensive demand intensity of the operation and maintenance task on network device resource consumption and safe operation.

[0134] Based on the task load coefficient threshold classification standard, all network device operation and maintenance tasks are divided into light-load tasks with low load and low resource requirements and heavy-load tasks with high load and high resource constraints. Stage task indicators are set for each, such as fault handling timeliness threshold and stage performance consumption control target. Overall task indicators, such as overall fault repair rate and overall energy consumption optimization ratio, are also set to clarify the management and control focus for different task types.

[0135] Based on the correlation between phase task indicators and overall task indicators and the expected change range during the operation and maintenance process, a state transition matrix is ​​constructed for each divided network device operation and maintenance task. The matrix dimensions cover key dimensions such as operation and maintenance stage, resource consumption level, and device operating status, and quantifies the transition probability and constraints between different states.

[0136] Based on the state transition matrix, the resource consumption and security evolution trend of the entire operation and maintenance process are deduced, and the standard deduction task process parameters of each network device operation and maintenance task after division are determined, including the preset resource occupation range, operation execution timing benchmark, security monitoring nodes, etc., and multiple core monitoring indicators such as energy consumption fluctuation, component load, and network transmission quality are selected.

[0137] Match the corresponding handling and control response sub-tasks for each monitoring indicator, such as component parameter adjustment, link bandwidth adaptation, program optimization and upgrade, etc., and clarify the scheduling resources required for each response sub-task, such as computing power resources, network bandwidth, and human operation and maintenance resources. Determine whether there are synchronous domain tasks that need to be executed synchronously through resource occupation conflict detection.

[0138] If there are synchronization domain task items, retrieve the network device control program corresponding to each divided network device operation and maintenance task, parse the operation instruction sequence and resource call logic in the program, extract the spatial interleaving features of the synchronization domain task items in the time dimension and execution logic, and identify possible resource competition conflict points and operational interference risks.

[0139] Based on the resource contention intensity and execution priority ranking of synchronous domain task items with spatial interleaved information characteristics, and combined with the resource constraints of network device operation, such as computing power limit and bandwidth limit, the resource requirements and handling and control priority weights of each network device operation and maintenance task are determined.

[0140] Based on the urgency of resource demand and the priority weight of handling and control, the network equipment operation and maintenance tasks in the light-load and heavy-load task types are sorted, and priority is given to ensuring the handling and control resources for high-load and high-security-risk tasks.

[0141] Based on the sorting results, the execution order of the handling and control is determined, and precise handling and control are implemented for each network device operation and maintenance task in sequence. Simultaneously, control feedback data is collected during the execution of each task, providing complete data support at the task dimension for subsequent dynamic correction of correlation and calibration of control accuracy.

[0142] In this embodiment, by combining task layering with precise control, the collaborative handling of network device faults and operational optimization are achieved. Relying on the adaptive association rule base of operating conditions, maintenance tasks are classified according to load coefficients, and targeted operational indicators and optimization goals are set. This makes the control more suitable for different task requirements. Heavy-load tasks focus on efficient handling under security constraints, while light-load tasks achieve ultimate optimization, improving the accuracy of operation and maintenance management. By predicting resource consumption and security evolution trends in advance and combining synchronous domain task item spatial interleaving analysis, operational conflicts and security interference caused by resource competition are avoided, ensuring the continuity and stability of control. Resources are allocated based on priority ranking, prioritizing high-load and high-risk tasks while balancing operation and maintenance efficiency and management security. Under the premise of ensuring the safe and stable operation of network devices, ineffective energy consumption is significantly reduced, resource utilization and the level of intelligent operation and maintenance management are improved, adapting to the network device operation and maintenance needs under multiple operating conditions.

[0143] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for fault early warning and rapid response handling based on network equipment operation and maintenance data, characterized in that, Includes the following steps: Visual and energy consumption perception: Real-time capture of data related to the hardware status, software operation status, and network topology link status of network devices; Synchronously collect target energy consumption data of key components of network equipment, establish the correlation between equipment status data and target energy consumption data of key components of network equipment, and generate a perception dataset; Fault correlation analysis: Obtain the sensing dataset, extract the target feature parameters from the sensing dataset and perform quantification and classification, and at the same time perform trend analysis on the target energy consumption data to extract abnormal energy consumption features; Analyze the correlation between target characteristic parameters and abnormal energy consumption characteristics, determine the type of potential fault and the target key components or link nodes that cause abnormal energy consumption and potential fault, and generate a fault correlation analysis report. Early warning and response decision generation: Based on the fault correlation analysis report, combined with the correlation degree between the target feature parameters and abnormal energy consumption characteristics corresponding to the perception dataset, the operation optimization space of the target key components is judged, and early warning prompts and response and handling plans adapted to different fault hazard types are generated, and the network equipment operation parameter benchmark is updated synchronously. Early warning response implementation and optimization: Obtain early warning prompts and response plans, updated operating parameter benchmarks, carry out targeted control and regulation of key components or link nodes, and simultaneously obtain control and regulation feedback data. Based on the control and regulation feedback data, dynamically correct the correlation of the sensing dataset and calibrate the control and regulation accuracy.

2. The fault early warning and rapid response handling method based on network equipment operation and maintenance data as described in claim 1, characterized in that, The visual and energy consumption perception also includes pre-setting perception nodes based on the operating conditions of network devices, deploying data acquisition devices in key components of the network device's core processor, power module, heat dissipation system, network interface card, and key link nodes of the network topology, and deploying energy consumption acquisition units in key components.

3. The fault early warning and rapid response handling method based on network equipment operation and maintenance data as described in claim 2, characterized in that, The specific process of generating a perceptual dataset includes: Based on the load characteristics and operational requirements of network devices under different operating conditions, dynamic parameter calibration is performed on the data acquisition devices at preset sensing nodes to obtain data acquisition adaptation parameters. Based on the energy consumption characteristics of each target key component, the range and accuracy of the pre-deployed energy consumption acquisition units are adapted and calibrated, and differentiated data acquisition frequencies and unified timestamps are set to obtain accurate energy consumption acquisition parameters. The data acquisition device captures various equipment status data in real time according to the adapted parameters, extracts the initial status characteristics, and the energy consumption acquisition unit collects the energy consumption data of each target key component in real time according to the precise parameters. By combining the energy consumption baseline range corresponding to the operating conditions, abnormal data is removed and effective energy consumption parameters are retained to form a preprocessed data set of state characteristics and energy consumption parameters; Based on the preprocessed dataset and combined with a unified timestamp, a one-to-one correspondence between status data and energy consumption data is established. At the same time, real-time operating condition tags of network devices are obtained, and a correlation model between status data, energy consumption data and real-time operating condition of network devices is established. Based on the correlation model, the inherent correlation patterns between state data and energy consumption data under different operating conditions are mined to generate an adaptive correlation rule base for operating conditions, thus forming structured correlation data. Based on structured associated data and an adaptive association rule base for operating conditions, the validity of the data and the rationality of the association are verified, and a perception dataset is generated based on the verification results.

4. The fault early warning and rapid response handling method based on network equipment operation and maintenance data as described in claim 3, characterized in that, The specific process of quantitative grading includes: Based on the working condition adaptive association rule base, target feature parameters such as CPU temperature exceeding the standard, abnormal memory usage, number of hard disk bad sectors, number of port connection interruptions, link packet loss rate, and protocol error frequency are extracted from the perception dataset under the corresponding working conditions. Based on network equipment security operation specifications and historical fault data, the security weights corresponding to each target characteristic parameter are determined, and a quantitative hierarchical evaluation matrix is ​​constructed. Cluster analysis is performed on the target feature parameters based on the quantitative grading evaluation matrix to obtain the grade interval corresponding to each target feature parameter and determine the safety level of each target feature parameter. The quantitative grading results are coupled with the abnormal energy consumption characteristics in the perception dataset for analysis. The correlation coefficient between the target feature parameters of each security level and the energy consumption anomaly type is output, and a mapping relationship table between the grading results and energy consumption anomalies is established. Historical operating data is acquired, and the accuracy of the classification results and the rationality of the correlation coefficients are verified based on the historical operating data. Based on the verification results, the probability of failure corresponding to the target characteristic parameters of different safety levels is determined in the mapping relationship table between the classification results and energy consumption anomalies.

5. The fault early warning and rapid response handling method based on network equipment operation and maintenance data as described in claim 4, characterized in that, Analyze the correlation between target characteristic parameters and abnormal energy consumption characteristics, specifically including: Obtain the mapping relationship table between the classification results and energy consumption anomalies, as well as the probability of failure corresponding to each safety level. Combine the working condition adaptive association rule library to extract the safety level and corresponding energy consumption anomaly type of the target feature parameters under the current working condition. Using the safety level of the target feature parameters as a reference sequence and the corresponding energy consumption anomaly type and failure probability as a comparison sequence, the correlation value between the target feature parameters of each safety level and different energy consumption anomaly types is calculated. Based on the network equipment security operation specifications, a correlation threshold is set, the calculated correlation value is compared with the correlation threshold, the correlation level is divided, and the priority of the corresponding energy consumption anomalies and faults caused by the target characteristic parameters of different levels is clarified. By combining the current operating condition labels and the operating condition adaptive association rule library, the association level is adjusted to adapt to the operating condition, and the corresponding relationship between the corrected target feature parameters, energy consumption anomaly types and association levels is output.

6. The fault early warning and rapid response handling method based on network equipment operation and maintenance data as described in claim 5, characterized in that, Using the safety level of the target characteristic parameters as a reference sequence, and the corresponding energy consumption anomaly type and failure probability as a comparison sequence, the correlation value between the target characteristic parameters of each safety level and different energy consumption anomaly types is calculated, including: Determine the probability of fault occurrence for each target feature parameter corresponding to the energy consumption anomaly type based on the reference sequence and comparison sequence; The probability of fault occurrence for the energy consumption anomaly type corresponding to the target feature parameter is normalized to obtain the normalized fault occurrence probability. The probability confidence function is used to process the normalized fault occurrence probability to obtain the fault probability coefficients under the energy consumption anomaly type corresponding to the target feature parameters; The actual occurrence rate of energy consumption anomaly types corresponding to the target feature parameters is retrieved from historical fault data, and an observation sequence is generated using the actual occurrence rate of energy consumption anomaly types corresponding to the target feature parameters. The similarity between the observed sequence and the comparison sequence corresponding to the target feature parameters is processed to obtain the similarity value between the observed sequence and the comparison sequence corresponding to the target feature parameters. The correlation between the target feature parameters of each safety level and different energy consumption anomaly types is calculated by combining the similarity values ​​between the observation sequences and comparison sequences corresponding to the target feature parameters with the failure probability coefficients under the energy consumption anomaly types corresponding to the target feature parameters.

7. The fault early warning and rapid response handling method based on network equipment operation and maintenance data as described in claim 6, characterized in that, The similarity between the observed sequence and the comparison sequence corresponding to the target feature parameters is processed to obtain the similarity value between the observed sequence and the comparison sequence corresponding to the target feature parameters, including: Retrieve the number of energy consumption anomaly types contained in the observation and comparison sequences corresponding to the target feature parameters; Retrieve the probability of failure and the actual percentage of occurrence for each type of energy consumption anomaly; The probability of failure and the actual occurrence rate corresponding to each type of energy consumption anomaly are compared to obtain the absolute difference between the probability of failure and the actual occurrence rate. The similarity value between the observation sequence and the comparison sequence corresponding to the target feature parameter is obtained by combining the absolute difference between the failure probability and the actual occurrence rate of each type of energy consumption anomaly contained in the observation sequence and the comparison sequence corresponding to the target feature parameter with the corresponding values ​​of the failure probability and the actual occurrence rate of each type of energy consumption anomaly.

8. The fault early warning and rapid response handling method based on network equipment operation and maintenance data as described in claim 5, characterized in that, The fault correlation analysis report includes: basic information, quantitative grading results, correlation analysis conclusions, fault hazard type, hazard level and corresponding target key component or link node information, and auxiliary decision-making data.

9. The fault early warning and rapid response handling method based on network equipment operation and maintenance data as described in claim 6, characterized in that, Early warning and response decision generation also includes: The decision-making criteria prioritize safety over energy optimization, and the basic fault handling control instructions are generated by combining the hazard level information in the fault correlation analysis report. Once scenarios with no or low-level potential hazards are identified, operation optimization control commands are generated based on the correspondence between target characteristic parameters, energy consumption anomaly types, and correlation levels, thereby adjusting the operating status of each target key component to the optimal range of the corresponding operating conditions. Establish instruction compatibility verification rules, cross-verify basic fault handling control instructions and operation optimization control instructions, eliminate conflicting control instructions, perform safety adaptation correction on boundary parameters, and finally generate an executable composite control instruction set.

10. The fault early warning and rapid response handling method based on network equipment operation and maintenance data as described in claim 9, characterized in that, The implementation and optimization of early warning response also includes: Obtain the network device operation and maintenance task description information corresponding to the response and handling plan, decompose the task description information into parameters, and extract the task target information, task type information, and task operation information. Based on task objective information, combined with task type information and task operation information, a load coefficient calculation model is constructed to quantitatively evaluate the task load coefficient of each network device operation and maintenance task. Based on the task load coefficient threshold classification standard, all network device operation and maintenance tasks are divided into light-load tasks and heavy-load tasks, and phase task indicators and overall task indicators are set respectively. Based on the correlation between phase task indicators and overall task indicators and the expected change range during the operation process, a state transition matrix is ​​constructed for each segmented network device operation and maintenance task. Based on the state transition matrix, the standard deduction task process parameters for each partitioned network device operation and maintenance task are determined, and core monitoring indicators are selected. Match the corresponding handling and control response sub-tasks for each monitoring indicator, clarify the scheduling resources required for each response sub-task, and determine whether there are any synchronization domain tasks that need to be executed synchronously through resource occupation conflict detection; If there are synchronization domain task items, retrieve the network device control program corresponding to each partitioned network device operation and maintenance task, and extract the spatial interleaving features of the synchronization domain task items in terms of time dimension and execution logic. Based on the resource contention intensity and execution priority ranking of synchronous domain task items with spatial interleaved information characteristics, and combined with the resource constraints of network device operation, the resource requirements and priority weights for handling and regulation of each network device operation and maintenance task are determined. Based on the urgency of loading space requirements and the priority weight of handling and control, the network equipment operation and maintenance tasks in the light-load operation task type and the heavy-load operation task type are sorted. The execution order of the handling and control is determined based on the sorting results. Precise handling and control are carried out sequentially for each network device operation and maintenance task, and control feedback data is collected simultaneously during the execution of each task.

Images