A method and system for processing of communication device health data

Abstract

The application discloses to the technical field of internet of things communication and device health monitoring, and particularly relates to a processing method and system for communication device health data. First, an adaptive collection strategy is constructed based on device portraits, while parameters are dynamically adjusted and multi-dimensional data is synchronously collected, and qualified original data sets are output after preliminary verification. Then, based on the qualified original data sets, heterogeneous standardization, anomaly identification, noise reduction and completion, and fusion and compression are sequentially completed, and high-quality standardized data sets are output. Based on this, the application performs multi-dimensional correlation analysis by combining device individualization benchmarks, same-type group benchmarks, and fault correlation characteristics, and has the advantages of realizing accurate evaluation of device health status, accurate positioning of hidden danger sources, and breaking away from dependence on artificial experience.

Description

Technical Field

[0001] This invention relates to the field of Internet of Things (IoT) communication and device health monitoring technology, specifically to a method and system for processing health data of communication devices. Background Technology

[0002] With the large-scale application of IoT technology, communication equipment, as the core carrier of IoT data transmission, directly determines the communication quality and service continuity of the IoT system through its operational stability. Currently, various IoT communication devices (such as outdoor base stations, industrial routers, and smart home gateways) are widely distributed in complex scenarios. Some devices are located in remote, unattended areas, high-temperature and high-humidity industrial environments, or harsh outdoor environments. During their operation, they generate massive amounts of health-related data, including hardware operating parameters (CPU utilization, memory usage, voltage and current), communication performance parameters (signal strength, transmission delay, packet loss rate), and environmental parameters (device surface temperature, ambient humidity, dust concentration).

[0003] The aforementioned health data is the core basis for judging the operating status of communication equipment and predicting potential faults. How to efficiently and accurately process this massive, multi-dimensional, and heterogeneous health data has become the key to improving the operation and maintenance efficiency of IoT communication equipment and reducing operation and maintenance costs.

[0004] In existing technologies, methods for processing health data of communication devices based on the Internet of Things (IoT) have the following specific problems: 1. Limited Health Analysis Dimensions, Reliance on Human Experience, and Low Prediction Accuracy: Existing methods often use single parameter thresholds to judge equipment health status, failing to combine historical operating data and benchmark data from similar equipment for multi-dimensional correlation analysis. Furthermore, warning thresholds are mostly manually set, unable to adapt to the dynamic adjustment needs arising from equipment aging and changing scenarios. For example, after three years of operation, a base station's hardware performance naturally ages, and the CPU utilization threshold should be lowered from 80% to 70%. However, existing methods still use the initial threshold, leading to frequent false alarms (a warning is issued when CPU utilization reaches 75%, when in reality it's normal operation after aging). Simultaneously, existing methods judge communication health solely based on packet loss rate, without considering transmission delay, signal strength, and other data, making it impossible to accurately pinpoint the cause of packet loss (whether it's due to hardware failure or environmental interference).

[0005] 2. Disconnect between early warning and feedback, lack of closed-loop management, and delayed operation and maintenance response: Existing methods can only provide early warnings of potential health hazards, but they do not link early warning information with equipment operation and maintenance processes or data update processes. There is no clear operation and maintenance guidance after an early warning, nor is the analysis model updated based on the operation and maintenance results. This renders early warning information ineffective and unable to quickly resolve equipment hazards. For example, if a router is warned of "excessive memory usage," the existing method only pushes the warning information to the operation and maintenance personnel, without clearly stating the specific reason for the excessive memory usage (whether it is excessive cache or hardware aging), nor providing targeted operation and maintenance suggestions (such as clearing the cache or replacing the memory). Furthermore, after the operation and maintenance personnel have handled the issue, the warning information is not simultaneously marked as processed, and subsequent analysis still uses the data before the processing, resulting in untimely model updates and the recurrence of similar hazards. Summary of the Invention

[0006] To address the aforementioned technical problems of limited health analysis dimensions and a disconnect between early warning and feedback, this invention provides the following technical solution: A method for processing health data of communication devices includes the following specific steps: S1, Differentiated health data collection: Based on device profiles, an adaptive collection strategy is built, while parameters are dynamically adjusted and multi-dimensional data is collected synchronously. After preliminary verification, a qualified raw dataset is output. S2, Intelligent data preprocessing: Based on the qualified raw dataset output by S1, heterogeneous standardization, anomaly identification, noise reduction and completion, and fusion compression are completed in sequence, and a high-quality standardized dataset is output. S3: Multi-dimensional health fusion analysis: Based on the high-quality standardized dataset output by S2, multi-dimensional analysis, trend prediction and hidden danger location are carried out in combination with the health benchmark model, and an equipment health status assessment report is output. S4, Closed-loop early warning and operation and maintenance linkage: Based on the equipment health status assessment report output by S3, hierarchical early warning, personalized operation and maintenance, process acceptance and model optimization are implemented in sequence to achieve closed-loop management and output relevant results.

[0007] As a preferred embodiment of the method for processing health data of communication equipment according to the present invention, the specific steps of S1 are as follows: S11, Device Profile Construction: Collect basic information of target communication devices through IoT terminal modules; classify and label the basic information to construct a unique profile for each device and output device-differentiated collection benchmarks; S12, Dynamic adjustment of acquisition strategy: Based on the device-differentiated acquisition benchmark output in S11, combined with the IoT transmission bandwidth usage and real-time device load data, dynamically adjust the acquisition parameters of each device; and after the adjustment is completed, output the adaptability acquisition parameters of each device. S13, Multi-source data synchronous acquisition: Based on the adaptive acquisition parameters output in S12, multi-dimensional health data of the target communication device are synchronously acquired through IoT sensor modules and device built-in monitoring units; and after the acquisition is completed, all data are summarized and the original acquisition dataset is output. S14, Preliminary verification of collected data: Based on the original collected dataset output in S13, perform preliminary verification on each data item; at the same time, mark the data that fails verification and temporarily store it, and summarize the data that passes verification and output the qualified original dataset.

[0008] As a preferred embodiment of the method for processing health data of communication equipment according to the present invention, the specific steps of S2 are as follows: S21, Heterogeneous Data Standardization: Based on the qualified original dataset output by S1, identify heterogeneous data in the dataset and use a hierarchical standardization algorithm for unified processing; at the same time, add dimension identifiers to each type of standardized data and output a dataset in a unified format. S22, Accurate identification of abnormal data: Based on the unified format dataset output in S21, combined with the device profile, abnormal data is accurately identified by threshold judgment and correlation verification, and a list of abnormal data is output; at the same time, the remaining normal data is summarized and a normal dataset is output. S23, Data Denoising and Completion: Based on the normal dataset and abnormal data details output in S22, the data is denoised and completed; and after the processing is completed, a complete preprocessed dataset is output. S24, Data Fusion and Compression: Based on the complete preprocessed dataset output in S23, a lightweight fusion and compression algorithm is used to fuse and compress multi-dimensional data; and after compression, a high-quality standardized dataset is output.

[0009] As a preferred embodiment of the method for processing health data of communication equipment according to the present invention, the specific steps of S3 are as follows: S31, Benchmark Database Construction: Based on device profiles, collect historical health data of target communication devices and health data of devices of the same model and in the same scenario, and perform statistical analysis on the collected data to extract the normal fluctuation range, aging trend pattern and fault correlation characteristics of data in each dimension; and combine the extracted features to construct a personalized health benchmark model for each device, and at the same time construct a group benchmark model for devices of the same type, and output a device health benchmark model containing personalized benchmarks and group benchmarks. S32, Multi-dimensional Correlation Analysis: Based on the equipment health benchmark model output by S31 and the high-quality standardized dataset output by S2, a multi-dimensional correlation analysis algorithm is used to quantitatively score the equipment health status; and after the analysis is completed, the quantitative score of the equipment health status and the score details of each dimension are output. S33, Dynamic prediction of health trends: Based on the quantitative score of device health status and detailed scores of each dimension output by S32, a time-series prediction algorithm is used to dynamically predict the future health status change trend of the device and output a trend prediction report. S34, Precise Location and Classification of Hidden Dangers: Based on the trend prediction report output in S33 and combined with the details of abnormal data, the potential hidden dangers and their root causes in the equipment are precisely located; at the same time, the hidden dangers are classified into Level 1, Level 2, and Level 3 according to their impact range, development speed, and degree of harm; and after the location is completed, the details of the hidden dangers are output. S35, Health Report Generation: Based on the equipment health status quantitative score output in S32, the trend prediction report output in S33, and the hidden danger details output in S34, combined with the equipment profile, an equipment health status assessment report is generated; and after generation, the equipment health status assessment report is output.

[0010] As a preferred embodiment of the method for processing health data of communication equipment according to the present invention, the specific steps of S4 are as follows: S41, Tiered early warning information push: Based on the equipment health status assessment report output by S3, a tiered early warning strategy is formulated according to the level of hidden danger; and after the push is completed, the response of operation and maintenance personnel is tracked, the response time and the responding personnel are recorded, and an early warning receipt is output; S42, Personalized generation of operation and maintenance solutions: Based on the early warning receipt output in S41 and combined with the device profile, a personalized operation and maintenance solution is generated for each device; and after the generation is completed, a targeted operation and maintenance guide is output. S43, Operation and Maintenance Process Tracking and Acceptance: Based on the targeted operation and maintenance guidelines output in S42, the operation and maintenance process is tracked in real time through the IoT terminal module; at the same time, after the operation and maintenance is completed, the operation and maintenance personnel upload photos of the operation and maintenance site and operation records, so as to automatically accept the operation and maintenance effect by combining the high-quality standardized dataset output in S2 and the equipment health status assessment report output in S3; and after the acceptance is completed, an operation and maintenance result report is output. S44, Data Update and Model Optimization: Based on the operation and maintenance result report output by S43, the equipment health status assessment report output by S3, and the high-quality standardized dataset output by S2, update and optimize the equipment health data and benchmark model; and after the update and optimization are completed, output the model update instruction and synchronize it to the IoT cloud platform and the acquisition terminal.

[0011] A system for processing health data of communication devices, comprising: The differentiated health data collection module builds an adaptive collection strategy based on device profiles, dynamically adjusts parameters and collects multi-dimensional data simultaneously, and outputs qualified raw datasets after preliminary verification. The intelligent data preprocessing module, based on the qualified raw dataset output by the differentiated health data acquisition module, sequentially completes heterogeneous standardization, anomaly identification, noise reduction and completion, and fusion compression, and outputs a high-quality standardized dataset. The multi-dimensional health fusion analysis module, based on the high-quality standardized dataset output by the intelligent data preprocessing module, combines the health benchmark model to carry out multi-dimensional analysis, trend prediction and hidden danger location, and outputs equipment health status assessment report; The closed-loop early warning and operation and maintenance linkage module, based on the equipment health status assessment report output by the multi-dimensional health fusion analysis module, sequentially implements graded early warning, personalized operation and maintenance, process acceptance and model optimization to achieve closed-loop management and output relevant results.

[0012] As a preferred embodiment of the health data processing system for communication equipment described in this invention, the differentiated health data acquisition module includes: The device profiling unit collects basic information about target communication devices through IoT terminal modules; it also classifies and labels the basic information to build a unique profile for each device and outputs differentiated data collection benchmarks for each device. The data acquisition strategy dynamic adjustment unit dynamically adjusts the acquisition parameters of each device based on the device-differentiated acquisition benchmark output by the device profile construction unit, combined with IoT transmission bandwidth usage and real-time device load data; and after the adjustment is completed, it outputs the adaptability acquisition parameters of each device. The multi-source data synchronous acquisition unit dynamically adjusts the adaptive acquisition parameters output by the unit based on the acquisition strategy, and synchronously acquires multi-dimensional health data of the target communication device through IoT sensor modules and device built-in monitoring units; and after the acquisition is completed, it summarizes all the data and outputs the original acquisition dataset. The preliminary data verification unit performs preliminary verification on each data item based on the original data set output by the multi-source data synchronous acquisition unit; at the same time, it marks the data that fails the verification and temporarily stores it, and summarizes the data that passes the verification and outputs the qualified original data set.

[0013] As a preferred embodiment of the health data processing system for communication equipment described in this invention, the intelligent data preprocessing module includes: The heterogeneous data standardization unit identifies heterogeneous data in the dataset based on the qualified raw dataset output by the differentiated health data collection module, and performs unified processing using a hierarchical standardization algorithm; at the same time, it adds dimension identifiers to each type of standardized data and outputs a dataset in a unified format. The abnormal data accurate identification unit, based on the unified format dataset output by the heterogeneous data standardization unit, combined with the device profile, uses threshold judgment plus correlation verification to accurately identify abnormal data and output abnormal data details; at the same time, it summarizes the remaining normal data and outputs the normal dataset. The data denoising and completion unit performs denoising and completion processing on the normal dataset and abnormal data details output by the abnormal data accurate identification unit; and outputs a complete preprocessed dataset after processing. The data fusion and compression unit, based on the complete preprocessed dataset output by the data denoising and completion unit, uses a lightweight fusion compression algorithm to fuse and compress multi-dimensional data; and after compression, it outputs a high-quality standardized dataset.

[0014] As a preferred embodiment of the health data processing system for communication equipment described in this invention, the multi-dimensional health fusion analysis module includes: The benchmark database construction unit collects historical health data of target communication devices and health data of devices of the same model and in the same scenario based on device profiles. It also performs statistical analysis on the collected data to extract the normal fluctuation range, aging trend pattern and fault correlation characteristics of each dimension of data. Based on the extracted features, it constructs a personalized health benchmark model for each device and a group benchmark model for devices of the same type. The output is a device health benchmark model containing personalized benchmarks and group benchmarks. The multi-dimensional correlation analysis unit, based on the equipment health benchmark model output by the benchmark database construction unit and the high-quality standardized dataset output by the intelligent data preprocessing module, uses a multi-dimensional correlation analysis algorithm to quantitatively score the equipment health status; and after the analysis is completed, it outputs the quantitative score of the equipment health status and the score details of each dimension. The health trend dynamic prediction unit, based on the equipment health status quantitative score and score details of each dimension output by the multi-dimensional correlation analysis unit, uses a time-series prediction algorithm to dynamically predict the future health status change trend of the equipment and outputs a trend prediction report. The hazard precise location and classification unit, based on the trend prediction report output by the health trend dynamic prediction unit and combined with abnormal data details, accurately locates potential equipment hazards and their root causes; at the same time, it classifies hazards into level one, level two, and level three according to their impact range, development speed, and degree of harm; and outputs hazard details after location is completed. The health report generation unit generates an equipment health status assessment report based on the equipment health status quantitative score output by the multi-dimensional correlation analysis unit, the trend prediction report output by the health trend dynamic prediction unit, and the hazard details output by the hazard precise location and level classification unit, combined with the equipment profile; and outputs the equipment health status assessment report after generation.

[0015] As a preferred embodiment of the communication equipment health data processing system described in this invention, the closed-loop early warning and operation and maintenance linkage module includes: The graded early warning information push unit, based on the equipment health status assessment report output by the multi-dimensional health fusion analysis module, formulates graded early warning strategies according to the level of hidden dangers; and after the push is completed, it tracks the response of operation and maintenance personnel, records the response time and the responding personnel, and outputs an early warning receipt. The personalized operation and maintenance plan generation unit generates a personalized operation and maintenance plan for each device based on the warning receipts output by the hierarchical warning information push unit and the device profile; and outputs a targeted operation and maintenance guide after the generation is completed. The operation and maintenance process tracking and acceptance unit, based on the targeted operation and maintenance guidelines output by the personalized operation and maintenance solution generation unit, tracks the operation and maintenance process in real time through the IoT terminal module. Simultaneously, after the operation and maintenance is completed, the operation and maintenance personnel upload photos of the operation and maintenance site and operation records. Combined with the high-quality standardized dataset output by the intelligent data preprocessing module and the equipment health status assessment report output by the multi-dimensional health fusion analysis module, the operation and maintenance effectiveness is automatically accepted. Furthermore, after acceptance, an operation and maintenance result report is output. The data update and model optimization unit updates and optimizes the equipment health data and benchmark model based on the operation and maintenance result report output by the operation and maintenance process tracking and acceptance unit, the equipment health status assessment report output by the multi-dimensional health fusion analysis module, and the high-quality standardized dataset output by the intelligent data preprocessing module; and after the update and optimization are completed, it outputs the model update instruction and synchronizes it to the IoT cloud platform and the acquisition terminal.

[0016] Compared with existing technologies: 1. By combining personalized equipment benchmarks, benchmarks of similar groups, and fault correlation characteristics, multi-dimensional correlation analysis is conducted, which has the advantages of achieving accurate assessment of equipment health status, accurate location of potential root causes, and eliminating reliance on human experience; 2. By linking early warning information, personalized operation and maintenance solutions, operation and maintenance acceptance, and model optimization, a closed-loop management and control mechanism is constructed, which has the advantages of achieving efficient early warning implementation, accurate operation and maintenance response, and continuous optimization of data processing. Attached Figure Description

[0017] Figure 1This is a schematic diagram of the overall framework of the present invention; Figure 2 This is a schematic diagram of the framework of the differentiated health data collection module of the present invention; Figure 3 This is a schematic diagram of the intelligent data preprocessing module framework of the present invention; Figure 4 This is a schematic diagram of the framework of the multi-dimensional health fusion analysis module of the present invention; Figure 5 This is a schematic diagram of the closed-loop early warning and operation and maintenance linkage module framework of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0019] This invention provides a method for processing health data of communication devices, comprising the following specific steps: S1, Differentiated health data collection: Based on device profiles, an adaptive collection strategy is built, while parameters are dynamically adjusted and multi-dimensional data is collected synchronously. After preliminary verification, a qualified raw dataset is output. The specific steps of S1 are as follows: S11, Device Profile Construction: Collect basic information of target communication devices through IoT terminal modules, including device type (base station, router, switch, etc.), device model, years of operation, core hardware parameters (CPU model, memory capacity, sensor type), deployment scenario (outdoor, indoor, industrial site, etc.), and core load requirements (average daily data transmission volume, number of concurrent connections); at the same time, classify and label the basic information to build a unique profile for each device, clarify the core health monitoring dimensions of each device (such as outdoor devices focusing on monitoring ambient temperature and humidity, and industrial devices focusing on monitoring voltage and current), and output differentiated collection benchmarks for each device; S12, Dynamic Adjustment of Acquisition Strategy: Based on the device-differentiated acquisition benchmark output in S11, combined with IoT transmission bandwidth occupancy and real-time device load data, the acquisition parameters of each device are dynamically adjusted; specifically: when the device load reaches a preset threshold (e.g., the number of concurrent connections exceeds 80% of the rated value), the acquisition frequency of key acquisition dimensions is increased by 50%, and the acquisition frequency of non-key acquisition dimensions is reduced by 30%; when the IoT transmission bandwidth occupancy rate exceeds 70%, the acquisition frequency of non-core devices is temporarily reduced to prioritize data acquisition of core devices (e.g., base stations); and after the adjustment is completed, the adaptability acquisition parameters of each device are output; S13, Multi-source data synchronous acquisition: Based on the adaptive acquisition parameters output in S12, multi-dimensional health data of the target communication device is synchronously acquired through IoT sensor modules and the device's built-in monitoring unit. This includes hardware operation data (CPU utilization, memory usage, voltage, current, fan speed), communication performance data (signal strength, transmission delay, packet loss rate, bandwidth utilization), and environmental correlation data (device surface temperature, ambient humidity, dust concentration). During the acquisition process, a timestamp, a unique device identifier (generated based on the device profile), and an acquisition dimension identifier are added to each data point to ensure data traceability. After acquisition, all data is aggregated and the original acquisition dataset is output. S14, Preliminary verification of collected data: Based on the original collected dataset output in S13, each data item is preliminarily verified. The verification includes: whether the data format conforms to the preset standard (e.g., whether the percentage format and numerical format are correct), whether the data exceeds the normal operating range of the equipment (e.g., whether the voltage data exceeds the rated voltage of the equipment by ±10%), and whether there are obvious transmission errors in the data (e.g., the value is 0 or empty). At the same time, the data that fails the verification is marked (indicating the error type) and temporarily stored (for subsequent anomaly analysis). The data that passes the verification are summarized and a qualified original dataset is output.

[0020] S2, Intelligent data preprocessing: Based on the qualified raw dataset output by S1, heterogeneous standardization, anomaly identification, noise reduction and completion, and fusion compression are completed in sequence, and a high-quality standardized dataset is output. The specific steps of S2 are as follows: S21, Heterogeneous Data Standardization: Based on the qualified original dataset output in S1, heterogeneous data (data of different formats and magnitudes) in the dataset are identified, and a hierarchical standardization algorithm is used for unified processing; specifically: percentage data (such as CPU utilization) is normalized and mapped to the [0,1] interval; numerical data (such as voltage and temperature) is standardized and converted into standardized values ​​with a mean of 0 and a variance of 1; decibel data (such as signal strength) is logarithmically transformed and uniformly converted into a numerical format that can be directly analyzed; at the same time, a dimension identifier is added to each type of standardized data to ensure data traceability and output a unified format dataset; S22, Accurate Identification of Abnormal Data: Based on the unified format dataset output in S21, combined with the equipment profile (normal operating parameter range of the equipment), abnormal data is accurately identified using threshold judgment and correlation verification. Specifically: First, based on the normal operating range in the equipment profile, data exceeding the threshold is filtered out and marked as suspected abnormal data; Second, combined with data of the same period and dimension, and data of related dimensions (such as suspected abnormal current data combined with voltage data and load data of the same period), it is verified to determine whether it is "abnormality caused by sensor failure", "abnormality caused by transmission interference" or "abnormality caused by actual equipment failure"; Third, the confirmed abnormal data is summarized, the abnormality type and the time of occurrence are marked, and the abnormal data details are output; at the same time, the remaining normal data is summarized and the normal dataset is output. S23, Data Denoising and Completion: Based on the normal dataset and abnormal data details output in S22, the data is denoised and completed. For minor fluctuations in the normal dataset (such as ±0.1℃ fluctuations in ambient temperature), a moving average algorithm is used for noise reduction while preserving the core trend of the data. For "missing data due to transmission interruption" in the abnormal data details, an interpolation algorithm is used to accurately complete the data by combining the corresponding data from similar devices in the same period and the historical data of the device in the same period (avoiding data distortion caused by simple filling). For "abnormal data caused by sensor failure", no completion is performed; only sensor failure information is marked for subsequent maintenance prompts. After processing, a complete preprocessed dataset is output. S24, Data Fusion and Compression: Based on the complete preprocessed dataset output in S23, a lightweight fusion compression algorithm is used to fuse and compress multi-dimensional data. Specifically, highly correlated dimensional data (such as CPU utilization and memory usage, signal strength and packet loss rate) are fused to extract core features (such as "high load - low signal") and reduce data redundancy. At the same time, the fused data is compressed to reduce the data volume (compression ratio controlled at 40%-60%) without losing core information, adapting to the narrow bandwidth transmission requirements of the Internet of Things. After compression, a high-quality standardized dataset is output.

[0021] S3: Multi-dimensional health fusion analysis: Based on the high-quality standardized dataset output by S2, multi-dimensional analysis, trend prediction and hidden danger location are carried out in combination with the health benchmark model, and an equipment health status assessment report is output. The specific steps of S3 are as follows: S31, Benchmark Database Construction: Based on device profiles, collect historical health data of target communication devices (preprocessed data from the past 1-3 years) and health data of devices of the same model and in the same scenario (retrieved from the IoT cloud database). Perform statistical analysis on the collected data to extract the normal fluctuation range, aging trend patterns (e.g., a 5% annual decrease in the normal CPU usage threshold) and fault correlation features (e.g., "temperature consistently above 60℃ + current fluctuation exceeding 10%" corresponds to hardware faults). Combine the extracted features to construct a personalized health benchmark model for each device, and simultaneously construct a group benchmark model for devices of the same type. Output a device health benchmark model containing personalized and group benchmarks. S32, Multi-dimensional Correlation Analysis: Based on the device health benchmark model output in S31 and the high-quality standardized dataset output in S2, a multi-dimensional correlation analysis algorithm is used to quantitatively score the device health status. Specifically, the analysis is broken down into three dimensions: hardware health, communication health, and environmental adaptability. Each dimension corresponds to multiple core indicators (e.g., hardware health corresponds to CPU utilization, memory usage, and voltage stability). Different weights are assigned to each indicator based on its deviation from the benchmark model (core indicators have higher weights than non-core indicators). The health score for each dimension is calculated, and then the overall device health score (out of 100) is calculated. A score ≥ 85 indicates a healthy state, 70 ≤ score < 85 indicates a sub-healthy state, and a score < 70 indicates an abnormal state. After the analysis is completed, the quantitative score of the device health status and detailed scores for each dimension are output. S33, Dynamic Health Trend Prediction: Based on the quantitative score of device health status and detailed scores of each dimension output in S32, a time-series prediction algorithm is used to dynamically predict the health status change trend of the device in the future preset period (e.g., 7 days, 30 days). Specifically, by combining the fluctuation pattern of scores of each dimension and the aging trend of the device, the direction of health score change (rising, stabilizing, falling) is predicted, potential hidden dangers are identified (e.g., the score of a certain dimension is slowly decreasing but has not reached the sub-health threshold, which is predicted to be a hidden aging danger), the trend deterioration nodes and risk levels are marked, and a trend prediction report is output. S34, Precise Location and Classification of Hidden Dangers: Based on the trend prediction report output in S33, combined with detailed abnormal data, the potential hidden dangers and root causes of the hidden dangers are precisely located. Specifically, for sub-healthy or abnormal equipment, dimensions and corresponding indicators with low scores are selected. Combined with the fault association feature library, the type of hidden danger (such as hardware aging, sensor failure, environmental interference, excessive load) is determined, and the root cause of the hidden danger is traced (such as "low signal strength score", which is determined to be "signal interference caused by heavy rain outdoors" based on environmental data, and "antenna aging aggravates interference" based on historical equipment data). At the same time, according to the scope of impact of the hidden danger (such as a single device or multiple devices in an area), the speed of development (such as rapid deterioration or slow accumulation), and the degree of harm (such as causing equipment downtime or only affecting transmission efficiency), the hidden dangers are classified into Level 1 (urgent), Level 2 (general), and Level 3 (minor). After the location is completed, the hidden danger details are output. S35, Health Report Generation: Based on the equipment health status quantitative score output in S32, the trend prediction report output in S33, and the hidden danger details output in S34, combined with the equipment profile, an equipment health status assessment report is generated. The report content includes: basic equipment information, current health score and performance in each dimension, details of potential hidden dangers, prediction of hidden danger development trends, and health status comparison (compared with its own historical average and the average level of similar equipment). The report adopts a visual format (including trend charts and scoring radar charts) for easy viewing by maintenance personnel. After generation, the equipment health status assessment report is output.

[0022] S4, Closed-loop early warning and operation and maintenance linkage: Based on the equipment health status assessment report output by S3, hierarchical early warning, personalized operation and maintenance, process acceptance and model optimization are implemented in sequence to achieve closed-loop management and output relevant results; The specific steps of S4 are as follows: S41, Tiered Early Warning Information Push: Based on the device health status assessment report output by S3, a tiered early warning strategy is formulated according to the level of potential hazards; Level 1 Hazard (Urgent): Early warning information is immediately pushed to maintenance personnel through three channels: IoT cloud platform, maintenance APP, and SMS, clearly specifying the warning content, hazard location, and emergency handling requirements, requiring a response within 1 hour; Level 2 Hazard (General): Early warning information is pushed through the maintenance APP and cloud platform, specifying the handling time limit (within 24 hours); Level 3 Hazard (Minor): The warning is marked on the cloud platform, and maintenance personnel are regularly reminded to pay attention; and after the push is completed, the response status of maintenance personnel is tracked, the response time and the responding personnel are recorded, and an early warning receipt is output; S42, Personalized Generation of Operation and Maintenance Plans: Based on the early warning feedback output in S41 and combined with the equipment profile, a personalized operation and maintenance plan is generated for each device. Specifically, based on the type and root cause of the hazard, combined with the equipment model, operating years, and deployment scenario, targeted operation and maintenance steps are formulated (e.g., for sensor failure hazards, specifying the steps of "disassembling the faulty sensor - replacing it with a sensor of the same model - calibrating the data acquisition accuracy"). At the same time, suggestions for the tools and consumables required for operation and maintenance are provided, and the operation and maintenance quality acceptance standards are clarified. After generation, a targeted operation and maintenance guide is output. S43, Operation and Maintenance Process Tracking and Acceptance: Based on the targeted operation and maintenance guidelines output in S42, the operation and maintenance process is tracked in real time through the IoT terminal module (such as the arrival time of operation and maintenance personnel and the execution status of operation steps); at the same time, after the operation and maintenance is completed, the operation and maintenance personnel upload photos of the operation and maintenance site and operation records, which, together with the high-quality standardized dataset output in S2 (data after operation and maintenance) and the equipment health status assessment report output in S3, automatically accept the operation and maintenance effect (for example, if the hidden danger is "excessive memory usage", the acceptance indicator is that the memory usage rate drops to the benchmark range after operation and maintenance and remains stable for more than 1 hour); after the acceptance is qualified, the hidden danger is marked as "resolved"; if the acceptance is unqualified, a second operation and maintenance prompt is pushed; and after the acceptance is completed, an operation and maintenance result report is output; S44, Data Update and Model Optimization: Based on the maintenance result report output in S43, the equipment health status assessment report output in S3, and the high-quality standardized dataset output in S2, update and optimize the equipment health data and benchmark model. Specifically, supplement the historical health data of the equipment with the data after maintenance and the results of handling potential problems, and update the equipment health benchmark model (such as adjusting health thresholds and adding new fault correlation features based on maintenance results). At the same time, optimize the collection strategy, preprocessing algorithm, and analysis algorithm based on the maintenance feedback of multiple devices. After the update and optimization are completed, output the model update instruction and synchronize it to the IoT cloud platform and collection terminal to achieve continuous optimization of subsequent data processing. At the same time, summarize the data of the entire processing process and output a closed-loop processing summary report for subsequent maintenance experience accumulation.

[0023] Please refer to a system for processing health data of communication devices. Figure 1 ,include: The differentiated health data collection module builds an adaptive collection strategy based on device profiles, dynamically adjusts parameters and collects multi-dimensional data simultaneously, and outputs qualified raw datasets after preliminary verification. The intelligent data preprocessing module, based on the qualified raw dataset output by the differentiated health data acquisition module, sequentially completes heterogeneous standardization, anomaly identification, noise reduction and completion, and fusion compression, and outputs a high-quality standardized dataset. The multi-dimensional health fusion analysis module, based on the high-quality standardized dataset output by the intelligent data preprocessing module, combines the health benchmark model to carry out multi-dimensional analysis, trend prediction and hidden danger location, and outputs equipment health status assessment report; The closed-loop early warning and operation and maintenance linkage module, based on the equipment health status assessment report output by the multi-dimensional health fusion analysis module, sequentially implements graded early warning, personalized operation and maintenance, process acceptance and model optimization to achieve closed-loop management and output relevant results.

[0024] Please see Figure 2 The differentiated health data collection module includes: The device profiling unit collects basic information about target communication devices through IoT terminal modules, including device type (base station, router, switch, etc.), device model, years of operation, core hardware parameters (CPU model, memory capacity, sensor type), deployment scenario (outdoor, indoor, industrial site, etc.), and core load requirements (average daily data transmission volume, number of concurrent connections). Simultaneously, it categorizes and labels the basic information to construct a unique profile for each device, clarifying the core health monitoring dimensions for each device (e.g., outdoor devices focus on monitoring ambient temperature and humidity, while industrial devices focus on monitoring voltage and current), and outputs differentiated data collection benchmarks for each device. The dynamic adjustment unit for the acquisition strategy dynamically adjusts the acquisition parameters of each device based on the device-differentiated acquisition benchmark output by the device profile construction unit, combined with IoT transmission bandwidth utilization and real-time device load data. Specifically, when the device load reaches a preset threshold (e.g., the number of concurrent connections exceeds 80% of the rated value), the acquisition frequency of key acquisition dimensions is increased by 50%, and the acquisition frequency of non-key acquisition dimensions is decreased by 30%. When the IoT transmission bandwidth utilization exceeds 70%, the acquisition frequency of non-core devices is temporarily reduced to prioritize data acquisition of core devices (e.g., base stations). After the adjustment is completed, the adaptation acquisition parameters for each device are output. The multi-source data synchronous acquisition unit dynamically adjusts the adaptive acquisition parameters output by the unit based on the acquisition strategy. Through IoT sensor modules and the device's built-in monitoring unit, it synchronously acquires multi-dimensional health data of the target communication device, including hardware operation data (CPU utilization, memory usage, voltage, current, fan speed), communication performance data (signal strength, transmission delay, packet loss rate, bandwidth utilization), and environmental correlation data (device surface temperature, ambient humidity, dust concentration). During the acquisition process, a timestamp, a unique device identifier (generated based on the device profile), and an acquisition dimension identifier are added to each data point to ensure data traceability. After acquisition, all data is aggregated and the original acquisition dataset is output. The preliminary data verification unit performs preliminary verification on each data item based on the original data set output by the multi-source data synchronous acquisition unit. The verification includes: whether the data format conforms to the preset standard (e.g., whether the percentage format and numerical format are correct), whether the data exceeds the normal operating range of the equipment (e.g., whether the voltage data exceeds the rated voltage of the equipment by ±10%), and whether there are obvious transmission errors in the data (e.g., the value is 0 or empty). At the same time, the data that fails the verification is marked (indicating the error type) and temporarily stored (for subsequent anomaly analysis). The data that passes the verification are summarized and a qualified original dataset is output.

[0025] Please see Figure 3 The intelligent data preprocessing module includes: The heterogeneous data standardization unit, based on the qualified raw dataset output by the differentiated health data acquisition module, identifies heterogeneous data (data of different formats and magnitudes) in the dataset and performs unified processing using a hierarchical standardization algorithm. Specifically, percentage-based data (such as CPU utilization) is normalized and mapped to the [0,1] interval; numerical data (such as voltage and temperature) is standardized and converted into standardized values ​​with a mean of 0 and a variance of 1; decibel-based data (such as signal strength) is logarithmically transformed and uniformly converted into a directly analyzable numerical format. Simultaneously, a dimension identifier is added to each type of standardized data to ensure data traceability and output a unified format dataset. The abnormal data accurate identification unit, based on the unified format dataset output by the heterogeneous data standardization unit, and combined with the equipment profile (the normal operating parameter range of the equipment), accurately identifies abnormal data using a threshold judgment plus correlation verification method. Specifically: First, based on the normal operating range in the equipment profile, data exceeding the threshold is filtered out and marked as suspected abnormal data; Second, it is verified by combining data of the same period and dimension, and data of related dimensions (such as suspected abnormal current data combined with voltage data and load data of the same period) to determine whether it is "anomaly caused by sensor failure", "anomaly caused by transmission interference" or "anomaly caused by actual equipment failure"; Third, the confirmed abnormal data is summarized, the anomaly type and the time of occurrence are marked, and the abnormal data details are output; at the same time, the remaining normal data is summarized and the normal dataset is output. The data denoising and completion unit performs denoising and completion processing on the normal dataset and abnormal data details output by the abnormal data accurate identification unit. For minor fluctuation noise in the normal dataset (such as ±0.1℃ fluctuation in ambient temperature), a moving average algorithm is used for denoising to preserve the core trend of data change. For "missing data due to transmission interruption" in the abnormal data details, an interpolation algorithm is used for accurate completion by combining the corresponding data of the same type of equipment in the same period and the historical data of the same period of the equipment (avoiding data distortion caused by simple filling). For "abnormal data caused by sensor failure", no completion is performed, only sensor failure information is marked for subsequent operation and maintenance prompts. After processing, a complete preprocessed dataset is output. The data fusion and compression unit, based on the complete preprocessed dataset output by the data denoising and completion unit, employs a lightweight fusion compression algorithm to fuse and compress multi-dimensional data. Specifically, it fuses highly correlated dimensional data (such as CPU utilization and memory usage, signal strength and packet loss rate) to extract core features (such as "high load - low signal") and reduce data redundancy. Simultaneously, it compresses the fused data to reduce its size (compression ratio controlled at 40%-60%) without losing core information, adapting to the narrow bandwidth transmission requirements of the Internet of Things. After compression, it outputs a high-quality standardized dataset.

[0026] Please see Figure 4 The multi-dimensional health fusion analysis module includes: The benchmark database construction unit, based on device profiles, collects historical health data (preprocessed data from the past 1-3 years) of target communication devices and health data of devices of the same model and in the same scenario (retrieved from an IoT cloud database). It then performs statistical analysis on the collected data, extracting normal fluctuation ranges, aging trend patterns (e.g., a 5% annual decrease in the normal CPU usage threshold), and fault correlation characteristics (e.g., "temperature consistently above 60℃ + current fluctuation exceeding 10%" corresponding to hardware faults). Combining these extracted features, it constructs a personalized health benchmark model for each device and a group benchmark model for similar devices, outputting a device health benchmark model containing both personalized and group benchmarks. The multi-dimensional correlation analysis unit, based on the device health benchmark model output by the benchmark database construction unit and the high-quality standardized dataset output by the intelligent data preprocessing module, uses a multi-dimensional correlation analysis algorithm to quantitatively score the device health status. Specifically, it analyzes the device health from three dimensions: hardware health, communication health, and environmental adaptability. Each dimension corresponds to multiple core indicators (e.g., hardware health corresponds to CPU utilization, memory usage, and voltage stability). Different weights are assigned to each indicator based on its deviation from the benchmark model (core indicators have higher weights than non-core indicators). The health score for each dimension is calculated, and then the overall device health score (out of 100) is calculated. A score ≥ 85 indicates a healthy state, 70 ≤ score < 85 indicates a sub-healthy state, and a score < 70 indicates an abnormal state. After the analysis, the unit outputs a quantitative score for the device health status and detailed scores for each dimension. The health trend dynamic prediction unit, based on the quantitative score of equipment health status and detailed scores of each dimension output by the multi-dimensional correlation analysis unit, uses a time-series prediction algorithm to dynamically predict the health status change trend of the equipment in the future preset period (e.g., 7 days, 30 days). Specifically, it combines the fluctuation pattern of scores of each dimension and the aging trend of the equipment to predict the direction of change of health score (upward, stable, downward), identify potential hidden dangers (e.g., a score of a certain dimension slowly declines but does not reach the sub-health threshold, which is predicted to be a hidden aging danger), mark the trend deterioration nodes and risk levels, and output a trend prediction report. The hazard precise location and classification unit, based on the trend prediction report output by the health trend dynamic prediction unit and combined with abnormal data details, accurately locates potential hazards and their root causes in equipment. Specifically, for sub-healthy or abnormal equipment, it filters out dimensions and corresponding indicators with low scores, and, combined with a fault association feature library, determines the hazard type (e.g., hardware aging, sensor failure, environmental interference, excessive load), and traces the root cause (e.g., "low signal strength score," determined by environmental data as "signal interference caused by heavy rain outdoors," and by historical equipment data as "antenna aging exacerbating interference"). Simultaneously, based on the hazard's impact range (e.g., single device, multiple devices within a region), development speed (e.g., rapid deterioration, slow accumulation), and severity (e.g., causing equipment downtime, only affecting transmission efficiency), the hazard is classified into Level 1 (urgent), Level 2 (general), and Level 3 (minor). After location is completed, a hazard detail is output. The health report generation unit generates an equipment health status assessment report based on the equipment health status quantitative score output by the multi-dimensional correlation analysis unit, the trend prediction report output by the health trend dynamic prediction unit, and the hazard details output by the hazard precise location and level classification unit, combined with the equipment profile. The report content includes: basic equipment information, current health score and performance in each dimension, potential hazard details, hazard development trend prediction, and health status comparison (compared with its own historical average and the average level of similar equipment). The report adopts a visual format (including trend charts and scoring radar charts) for easy viewing by maintenance personnel. After generation, the equipment health status assessment report is output.

[0027] Please see Figure 5 The closed-loop early warning and operation and maintenance linkage module includes: The tiered early warning information push unit, based on the equipment health status assessment report output by the multi-dimensional health fusion analysis module, formulates tiered early warning strategies according to the level of potential hazards: Level 1 hazard (urgent): Early warning information is immediately pushed to maintenance personnel through three channels: IoT cloud platform, maintenance APP, and SMS, clearly specifying the warning content, hazard location, and emergency handling requirements, requiring a response within 1 hour; Level 2 hazard (general): Early warning information is pushed through the maintenance APP and cloud platform, specifying the handling time limit (within 24 hours); Level 3 hazard (minor): The early warning is marked on the cloud platform, and maintenance personnel are regularly reminded to pay attention; and after the push is completed, the response status of maintenance personnel is tracked, the response time and the responding personnel are recorded, and an early warning receipt is output; The personalized operation and maintenance plan generation unit generates a personalized operation and maintenance plan for each device based on the early warning receipts output by the hierarchical early warning information push unit and the device profile. Specifically, it formulates targeted operation and maintenance steps based on the type and root cause of the hidden danger, combined with the device model, years of operation, and deployment scenario (e.g., for sensor failure hidden dangers, it specifies the steps of "disassembling the faulty sensor - replacing it with a sensor of the same model - calibrating the data acquisition accuracy"). At the same time, it provides suggestions for the tools and consumables required for operation and maintenance, and clarifies the operation and maintenance quality acceptance standards. After the generation is completed, it outputs a targeted operation and maintenance guide. The operation and maintenance process tracking and acceptance unit, based on the targeted operation and maintenance guidelines output by the personalized operation and maintenance solution generation unit, tracks the operation and maintenance process in real time through the IoT terminal module (such as the arrival time of operation and maintenance personnel and the execution status of operation steps). Simultaneously, after the operation and maintenance is completed, the operation and maintenance personnel upload photos of the operation and maintenance site and operation records. Combined with the high-quality standardized dataset (post-operation and maintenance data) output by the intelligent data preprocessing module and the equipment health status assessment report output by the multi-dimensional health fusion analysis module, the operation and maintenance effect is automatically accepted (e.g., if the hidden danger is "excessive memory usage," the acceptance indicator is that the memory usage rate drops to the benchmark range after operation and maintenance and remains stable for more than 1 hour). After acceptance is passed, the hidden danger is marked as "resolved"; if acceptance fails, a secondary operation and maintenance prompt is pushed; and after acceptance is completed, an operation and maintenance result report is output. The data update and model optimization unit updates and optimizes equipment health data and benchmark models based on the operation and maintenance result report output by the operation and maintenance process tracking and acceptance unit, the equipment health status assessment report output by the multi-dimensional health fusion analysis module, and the high-quality standardized dataset output by the intelligent data preprocessing module. Specifically, it supplements the historical health data of the equipment with the data after operation and maintenance and the results of handling potential hazards, and updates the equipment health benchmark model (such as adjusting health thresholds and adding new fault correlation features based on operation and maintenance results). At the same time, it optimizes the collection strategy, preprocessing algorithm, and analysis algorithm based on the operation and maintenance feedback of multiple devices. After the update and optimization are completed, it outputs the model update instruction and synchronizes it to the IoT cloud platform and collection terminal to achieve continuous optimization of subsequent data processing. At the same time, it summarizes the data of the entire processing process and outputs a closed-loop processing summary report for subsequent operation and maintenance experience accumulation.

[0028] Although the present invention has been described above with reference to embodiments, various modifications can be made and components can be replaced with equivalents without departing from the scope of the invention. In particular, as long as there is no structural conflict, the features in the disclosed embodiments can be combined with each other in any manner. The lack of an exhaustive description of these combinations in this specification is merely for the sake of brevity and resource conservation. Therefore, the present invention is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A method for processing health data of communication equipment, characterized in that, The specific steps are as follows: S1, Differentiated health data collection: Based on device profiles, an adaptive collection strategy is built, while parameters are dynamically adjusted and multi-dimensional data is collected synchronously. After preliminary verification, a qualified raw dataset is output. S2, Intelligent data preprocessing: Based on the qualified raw dataset output by S1, heterogeneous standardization, anomaly identification, noise reduction and completion, and fusion compression are completed in sequence, and a high-quality standardized dataset is output. S3: Multi-dimensional health fusion analysis: Based on the high-quality standardized dataset output by S2, multi-dimensional analysis, trend prediction and hidden danger location are carried out in combination with the health benchmark model, and an equipment health status assessment report is output. S4, Closed-loop early warning and operation and maintenance linkage: Based on the equipment health status assessment report output by S3, hierarchical early warning, personalized operation and maintenance, process acceptance and model optimization are implemented in sequence to achieve closed-loop management and output relevant results.

2. The method for processing health data of communication equipment according to claim 1, characterized in that, The specific steps of S1 are as follows: S11, Device Profile Construction: Collect basic information of target communication devices through IoT terminal modules; classify and label the basic information to construct a unique profile for each device and output device-differentiated collection benchmarks; S12, Dynamic adjustment of acquisition strategy: Based on the device-differentiated acquisition benchmark output in S11, combined with the IoT transmission bandwidth usage and real-time load data of the devices, the acquisition parameters of each device are dynamically adjusted. Furthermore, after the adjustment is completed, the compatibility acquisition parameters for each device are output; S13, Multi-source data synchronous acquisition: Based on the adaptive acquisition parameters output in S12, multi-dimensional health data of the target communication device are synchronously acquired through IoT sensor modules and device built-in monitoring units; Furthermore, after the data collection is completed, all data is aggregated and the original collected dataset is output. S14, Preliminary verification of collected data: Based on the original collected dataset output in S13, perform preliminary verification on each data item; at the same time, mark the data that fails verification and temporarily store it, and summarize the data that passes verification and output the qualified original dataset.

3. The method for processing health data of communication equipment according to claim 1, characterized in that, The specific steps of S2 are as follows: S21, Heterogeneous Data Standardization: Based on the qualified original dataset output by S1, identify heterogeneous data in the dataset and use a hierarchical standardization algorithm for unified processing; at the same time, add dimension identifiers to each type of standardized data and output a dataset in a unified format. S22, Accurate identification of abnormal data: Based on the unified format dataset output in S21, combined with the device profile, abnormal data is accurately identified by threshold judgment and correlation verification, and a list of abnormal data is output; at the same time, the remaining normal data is summarized and a normal dataset is output. S23, Data Denoising and Completion: Based on the normal dataset and abnormal data details output in S22, the data is denoised and completed. Furthermore, after processing is complete, the entire preprocessed dataset is output; S24, Data Fusion and Compression: Based on the complete preprocessed dataset output in S23, a lightweight fusion and compression algorithm is used to fuse and compress multi-dimensional data; and after compression, a high-quality standardized dataset is output.

4. The method for processing health data of communication equipment according to claim 1, characterized in that, The specific steps of S3 are as follows: S31, Baseline Database Construction: Based on device profiles, collect historical health data of target communication devices and health data of devices of the same model and in the same scenario, and perform statistical analysis on the collected data to extract the normal fluctuation range, aging trend patterns and fault correlation characteristics of data in each dimension. Based on the extracted features, a personalized health benchmark model is constructed for each device, and a group benchmark model for devices of the same type is also constructed. The output is a device health benchmark model containing personalized and group benchmarks. S32, Multi-dimensional correlation analysis: Based on the equipment health benchmark model output by S31 and the high-quality standardized dataset output by S2, a multi-dimensional correlation analysis algorithm is used to quantitatively score the equipment health status. Furthermore, after the analysis is completed, the output includes a quantitative score of the device's health status and detailed scores for each dimension; S33, Dynamic prediction of health trends: Based on the quantitative score of device health status and detailed scores of each dimension output by S32, a time-series prediction algorithm is used to dynamically predict the future health status change trend of the device and output a trend prediction report. S34, Precise Location and Classification of Hidden Dangers: Based on the trend prediction report output in S33 and combined with the details of abnormal data, the potential hidden dangers and their root causes in the equipment are precisely located; at the same time, the hidden dangers are classified into Level 1, Level 2, and Level 3 according to their impact range, development speed, and degree of harm; and after the location is completed, the details of the hidden dangers are output. S35, Health Report Generation: Based on the equipment health status quantitative score output in S32, the trend prediction report output in S33, and the hidden danger details output in S34, combined with the equipment profile, an equipment health status assessment report is generated; and after generation, the equipment health status assessment report is output.

5. The method for processing health data of communication equipment according to claim 1, characterized in that, The specific steps of S4 are as follows: S41, Tiered early warning information push: Based on the equipment health status assessment report output by S3, a tiered early warning strategy is formulated according to the level of hidden danger; and after the push is completed, the response of operation and maintenance personnel is tracked, the response time and the responding personnel are recorded, and an early warning receipt is output; S42, Personalized generation of operation and maintenance plan: Based on the early warning receipt output in S41, combined with the device profile, a personalized operation and maintenance plan is generated for each device. Furthermore, once generated, a targeted operation and maintenance guide will be output; S43, Operation and Maintenance Process Tracking and Acceptance: Based on the targeted operation and maintenance guidelines output in S42, the operation and maintenance process is tracked in real time through the IoT terminal module; Simultaneously, after the operation and maintenance is completed, the operation and maintenance personnel upload photos of the operation and maintenance site and operation records, so as to automatically accept the operation and maintenance effect by combining the high-quality standardized dataset output by S2 and the equipment health status assessment report output by S3. Furthermore, an operation and maintenance result report will be output after acceptance. S44, Data Update and Model Optimization: Based on the operation and maintenance result report output by S43, the equipment health status assessment report output by S3, and the high-quality standardized dataset output by S2, update and optimize the equipment health data and benchmark model; and after the update and optimization are completed, output the model update instruction and synchronize it to the IoT cloud platform and the acquisition terminal.

6. A system for processing health data of communication equipment, characterized in that, include: The differentiated health data collection module builds an adaptive collection strategy based on device profiles, dynamically adjusts parameters and collects multi-dimensional data simultaneously, and outputs qualified raw datasets after preliminary verification. The intelligent data preprocessing module, based on the qualified raw dataset output by the differentiated health data acquisition module, sequentially completes heterogeneous standardization, anomaly identification, noise reduction and completion, and fusion compression, and outputs a high-quality standardized dataset. The multi-dimensional health fusion analysis module, based on the high-quality standardized dataset output by the intelligent data preprocessing module, combines the health benchmark model to carry out multi-dimensional analysis, trend prediction and hidden danger location, and outputs equipment health status assessment report; The closed-loop early warning and operation and maintenance linkage module, based on the equipment health status assessment report output by the multi-dimensional health fusion analysis module, sequentially implements graded early warning, personalized operation and maintenance, process acceptance and model optimization to achieve closed-loop management and output relevant results.

7. A system for processing health data of communication equipment according to claim 6, characterized in that, The differentiated health data collection module includes: The device profiling unit collects basic information about target communication devices through IoT terminal modules; it also classifies and labels the basic information to build a unique profile for each device and outputs differentiated data collection benchmarks for each device. The data acquisition strategy dynamic adjustment unit dynamically adjusts the acquisition parameters of each device based on the device-differentiated acquisition benchmark output by the device profile construction unit, combined with IoT transmission bandwidth usage and real-time device load data; and after the adjustment is completed, it outputs the adaptability acquisition parameters of each device. The multi-source data synchronous acquisition unit dynamically adjusts the adaptive acquisition parameters output by the unit based on the acquisition strategy, and synchronously acquires multi-dimensional health data of the target communication device through IoT sensor modules and device built-in monitoring units; and after the acquisition is completed, it summarizes all the data and outputs the original acquisition dataset. The preliminary data verification unit performs preliminary verification on each data item based on the original data set output by the multi-source data synchronous acquisition unit; at the same time, it marks the data that fails the verification and temporarily retains it, and summarizes the data that passes the verification and outputs the qualified original data set.

8. A system for processing health data of communication equipment according to claim 6, characterized in that, The intelligent data preprocessing module includes: The heterogeneous data standardization unit identifies heterogeneous data in the dataset based on the qualified raw dataset output by the differentiated health data collection module, and performs unified processing using a hierarchical standardization algorithm; at the same time, it adds dimension identifiers to each type of standardized data and outputs a dataset in a unified format. The abnormal data accurate identification unit, based on the unified format dataset output by the heterogeneous data standardization unit, combined with the device profile, uses threshold judgment plus correlation verification to accurately identify abnormal data and output abnormal data details; at the same time, it summarizes the remaining normal data and outputs the normal dataset. The data denoising and completion unit performs denoising and completion processing on the normal dataset and abnormal data details output by the abnormal data accurate identification unit; and outputs a complete preprocessed dataset after processing. The data fusion and compression unit, based on the complete preprocessed dataset output by the data denoising and completion unit, uses a lightweight fusion compression algorithm to fuse and compress multi-dimensional data; and after compression, it outputs a high-quality standardized dataset.

9. A system for processing health data of communication equipment according to claim 6, characterized in that, The multi-dimensional health fusion analysis module includes: The benchmark database construction unit collects historical health data of target communication devices and health data of devices of the same model and in the same scenario based on device profiles. It also performs statistical analysis on the collected data to extract the normal fluctuation range, aging trend pattern and fault correlation characteristics of each dimension of data. Based on the extracted features, it constructs a personalized health benchmark model for each device and a group benchmark model for devices of the same type. The output is a device health benchmark model containing personalized benchmarks and group benchmarks. The multi-dimensional correlation analysis unit, based on the equipment health benchmark model output by the benchmark database construction unit and the high-quality standardized dataset output by the intelligent data preprocessing module, uses a multi-dimensional correlation analysis algorithm to quantitatively score the equipment health status; and after the analysis is completed, it outputs the quantitative score of the equipment health status and the score details of each dimension. The health trend dynamic prediction unit, based on the equipment health status quantitative score and score details of each dimension output by the multi-dimensional correlation analysis unit, uses a time-series prediction algorithm to dynamically predict the future health status change trend of the equipment and outputs a trend prediction report. The hazard precise location and classification unit, based on the trend prediction report output by the health trend dynamic prediction unit and combined with abnormal data details, accurately locates potential equipment hazards and their root causes; at the same time, it classifies hazards into level one, level two, and level three according to their impact range, development speed, and degree of harm; and outputs hazard details after location is completed. The health report generation unit generates an equipment health status assessment report based on the equipment health status quantitative score output by the multi-dimensional correlation analysis unit, the trend prediction report output by the health trend dynamic prediction unit, and the hazard details output by the hazard precise location and level classification unit, combined with the equipment profile; and outputs the equipment health status assessment report after generation.

10. A system for processing health data of communication equipment according to claim 6, characterized in that, The closed-loop early warning and operation and maintenance linkage module includes: The graded early warning information push unit, based on the equipment health status assessment report output by the multi-dimensional health fusion analysis module, formulates graded early warning strategies according to the level of hidden dangers; and after the push is completed, it tracks the response of operation and maintenance personnel, records the response time and the responding personnel, and outputs an early warning receipt. The personalized operation and maintenance plan generation unit generates a personalized operation and maintenance plan for each device based on the warning receipts output by the hierarchical warning information push unit and the device profile; and outputs a targeted operation and maintenance guide after the generation is completed. The operation and maintenance process tracking and acceptance unit, based on the targeted operation and maintenance guidelines output by the personalized operation and maintenance solution generation unit, tracks the operation and maintenance process in real time through the IoT terminal module. Simultaneously, after the operation and maintenance is completed, the operation and maintenance personnel upload photos of the operation and maintenance site and operation records. Combined with the high-quality standardized dataset output by the intelligent data preprocessing module and the equipment health status assessment report output by the multi-dimensional health fusion analysis module, the operation and maintenance effectiveness is automatically accepted. Furthermore, after acceptance, an operation and maintenance result report is output. The data update and model optimization unit updates and optimizes the equipment health data and benchmark model based on the operation and maintenance result report output by the operation and maintenance process tracking and acceptance unit, the equipment health status assessment report output by the multi-dimensional health fusion analysis module, and the high-quality standardized dataset output by the intelligent data preprocessing module; and after the update and optimization are completed, it outputs the model update instruction and synchronizes it to the IoT cloud platform and the acquisition terminal.

Images