Super equipment fault detection method and system based on multi-dimensional data analysis

Abstract

The present invention discloses a super equipment fault detection method and system based on multi-dimensional data analysis, the method comprising: obtaining historical multi-dimensional data of the super equipment, and pre-processing the historical multi-dimensional data to obtain multi-dimensional index data; obtaining multi-dimensional data of faults in the multi-dimensional index data, and training a preset deep learning model according to the multi-dimensional data of faults to obtain a fault analysis model; obtaining current multi-dimensional fault data, inputting the current multi-dimensional fault data into the fault analysis model, and outputting the current fault analysis result of the super equipment. The present invention obtains multi-dimensional data of each sub-equipment in the super equipment, predicts and identifies abnormal data in the multi-dimensional data and abnormal states of the equipment, and provides solutions and suggestions for abnormal states, thereby effectively improving the operation and maintenance efficiency when a fault occurs; at the same time, it also ensures the stability of the normal operation of the super equipment.

Description

Technical Field

[0001] The present invention relates to the field of equipment detection technology, and in particular to a super equipment fault detection method, system, terminal and computer-readable storage medium based on multi-dimensional data analysis. Background Art

[0002] With the rapid development of Internet of Things technology, more and more devices are connected through the network to form distributed systems. Especially in the fields of smart home, smart office and Internet of Vehicles, multiple independent devices are connected through the operating system to form a super device, realizing resource sharing and functional coordination.

[0003] However, the management and operation of super devices face many challenges, especially when the number of devices is large and the types are diverse. It is impossible to monitor the health status of each sub-device in the super device in real time and handle abnormal conditions, which affects the normal operation of the super device.

[0004] Therefore, the prior art still needs to be improved and developed. Summary of the invention

[0005] The main purpose of the present invention is to provide a super device fault detection method, system, terminal and computer-readable storage medium based on multi-dimensional data analysis, aiming to solve the problem in the prior art that when there are a large number of devices and various types, the health status of each sub-device in the super device cannot be monitored in real time and the abnormal status that occurs cannot be processed, thereby affecting the normal operation of the super device.

[0006] To achieve the above object, the present invention provides a super device fault detection method based on multi-dimensional data analysis, and the super device fault detection method based on multi-dimensional data analysis comprises the following steps:

[0007] Acquire historical multi-dimensional data of the super device, and pre-process the historical multi-dimensional data to obtain multi-dimensional indicator data;

[0008] Acquire multi-dimensional fault data from the multi-dimensional indicator data, and train a preset deep learning model according to the multi-dimensional fault data to obtain a fault analysis model;

[0009] Acquire current multi-dimensional fault data, input the current multi-dimensional fault data into the fault analysis model, and output the current fault analysis result of the super device.

[0010] Optionally, in the super equipment fault detection method based on multi-dimensional data analysis, the historical multi-dimensional data includes historical equipment basic indicator data, historical key indicator event data and historical log data;

[0011] The obtaining of historical multi-dimensional data of the super device and preprocessing of the historical multi-dimensional data to obtain multi-dimensional indicator data specifically includes:

[0012] Acquire historical equipment basic indicator data of the super equipment, and perform data cleaning and normalization processing on the historical equipment basic indicator data to obtain equipment basic indicator data;

[0013] Acquire historical key indicator event data of the super device, and perform first data classification processing and first data filtering processing on the historical key indicator event data to obtain key indicator event data;

[0014] Acquire historical log data of the super device, and perform second data classification processing and second data filtering processing on the historical log data to obtain log data;

[0015] Among them, the multi-dimensional indicator data includes the basic indicator data of the equipment, the key indicator event data and the log data.

[0016] Optionally, the super equipment fault detection method based on multi-dimensional data analysis, wherein the multi-dimensional fault data in the multi-dimensional indicator data is obtained, and a preset deep learning model is trained according to the multi-dimensional fault data to obtain a fault analysis model, further comprising:

[0017] Aggregate the basic indicator data of the device, the key indicator event data and the log data to obtain aggregated multi-dimensional indicator data;

[0018] The aggregated multi-dimensional indicator data is subjected to time series conversion processing to obtain multi-dimensional time series data.

[0019] Optionally, the super equipment fault detection method based on multi-dimensional data analysis, wherein the acquiring of multi-dimensional fault data in the multi-dimensional indicator data, and training of a preset deep learning model according to the multi-dimensional fault data to obtain a fault analysis model, specifically includes:

[0020] Obtain a preset baseline model, and input the multidimensional time series data into the preset baseline model, and use an anomaly detection algorithm to identify abnormal events on the multidimensional time series data through the preset baseline model to obtain abnormal event data in the multidimensional time series data;

[0021] Acquire a preset fault mode library, and perform fault search in the preset fault mode library according to the abnormal event data to obtain multi-dimensional fault data corresponding to the abnormal event data;

[0022] Reprocessing the fault multi-dimensional data to obtain a fault multi-dimensional training set, wherein the reprocessing includes data cleaning processing, feature extraction processing and standardization processing;

[0023] A preset deep learning model is determined, and a cross-validation method and a hyperparameter tuning method are used to train the preset deep learning model according to the fault multi-dimensional training set to obtain a fault analysis model.

[0024] Optionally, in the super equipment fault detection method based on multi-dimensional data analysis, the preset deep learning model includes a convolutional neural network, a recursive neural network and a long short-term memory neural network.

[0025] Optionally, the super device fault detection method based on multi-dimensional data analysis, wherein the obtaining of current multi-dimensional fault data, inputting the current multi-dimensional fault data into the fault analysis model, and outputting the current fault analysis result of the super device, specifically includes:

[0026] Acquire current multi-dimensional fault data, and input the current multi-dimensional fault data into the fault analysis model;

[0027] The fault analysis model is used to perform fault analysis processing on the current multi-dimensional fault data to obtain a fault analysis result, wherein the fault analysis result includes a fault cause and a fault impact range.

[0028] Optionally, the super device fault detection method based on multi-dimensional data analysis, wherein the current multi-dimensional fault data is obtained, the current multi-dimensional fault data is input into the fault analysis model, and the fault analysis result of the super device is output, and then further includes:

[0029] A fault recovery suggestion library is obtained, and a search is performed in the fault recovery suggestion library according to the fault cause and the fault impact range to obtain a target fault recovery suggestion.

[0030] In addition, to achieve the above-mentioned purpose, the present invention further provides a super equipment fault detection system based on multi-dimensional data analysis, wherein the super equipment fault detection system based on multi-dimensional data analysis includes:

[0031] A data acquisition module is used to acquire historical multi-dimensional data of the super device and pre-process the historical multi-dimensional data to obtain multi-dimensional indicator data;

[0032] A model training module, used to obtain multi-dimensional fault data in the multi-dimensional indicator data, and train a preset deep learning model according to the multi-dimensional fault data to obtain a fault analysis model;

[0033] The model analysis module is used to obtain current multi-dimensional fault data, input the current multi-dimensional fault data into the fault analysis model, and output the current fault analysis result of the super device.

[0034] In addition, to achieve the above-mentioned purpose, the present invention also provides a terminal, wherein the terminal includes: a memory, a processor, and a super device fault detection program based on multi-dimensional data analysis stored in the memory and executable on the processor, wherein the super device fault detection program based on multi-dimensional data analysis implements the steps of the super device fault detection method based on multi-dimensional data analysis as described above when executed by the processor.

[0035] In addition, to achieve the above-mentioned purpose, the present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a super device fault detection program based on multi-dimensional data analysis, and when the super device fault detection program based on multi-dimensional data analysis is executed by a processor, the steps of the super device fault detection method based on multi-dimensional data analysis as described above are implemented.

[0036] In the present invention, the historical multi-dimensional data of the super device is obtained, and the historical multi-dimensional data is preprocessed to obtain multi-dimensional indicator data; the multi-dimensional data of faults in the multi-dimensional indicator data is obtained, and the preset deep learning model is trained according to the multi-dimensional data of faults to obtain a fault analysis model; the current multi-dimensional fault data is obtained, and the current multi-dimensional fault data is input into the fault analysis model, and the current fault analysis result of the super device is output. The present invention obtains the multi-dimensional data of each sub-device in the super device, and combines AI technology to perform intelligent analysis on the multi-dimensional data, so as to quickly predict and identify abnormal data in the multi-dimensional data and abnormal conditions of the equipment, and provide accurate fault causes and recovery suggestions, which greatly improves the response speed to the equipment status, so that the operation and maintenance personnel can quickly obtain the key indicators of the equipment operation, reduce the need for manual inspection, and thus significantly improve the efficiency of operation and maintenance; at the same time, it also effectively improves the stability of the operation of the super device. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is a flow chart of a preferred embodiment of the super equipment fault detection method based on multi-dimensional data analysis of the present invention;

[0038] Figure 2 It is a schematic diagram of the system architecture of a preferred embodiment of the super equipment fault detection method based on multi-dimensional data analysis of the present invention;

[0039] Figure 3 It is a schematic diagram of the data collection process of a preferred embodiment of the super equipment fault detection method based on multi-dimensional data analysis of the present invention;

[0040] Figure 4 It is a schematic diagram of the data analysis process of a preferred embodiment of the super equipment fault detection method based on multi-dimensional data analysis of the present invention;

[0041] Figure 5 It is a schematic diagram of an event trend analysis process of a preferred embodiment of a super equipment fault detection method based on multi-dimensional data analysis of the present invention;

[0042] Figure 6 It is a schematic diagram of the AI ​​fault analysis and recovery process of a preferred embodiment of the super equipment fault detection method based on multi-dimensional data analysis of the present invention;

[0043] Figure 7 It is a structural diagram of a preferred embodiment of a super equipment fault detection system based on multi-dimensional data analysis of the present invention;

[0044] Figure 8 It is a structural diagram of a preferred embodiment of the terminal of the present invention. DETAILED DESCRIPTION

[0045] In order to make the purpose, technical solution and advantages of the present invention clearer and more specific, the present invention is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0046] With the rapid development of Internet of Things technology, more and more devices are connected through the network to form distributed systems. Especially in the fields of smart home, smart office and Internet of Vehicles, multiple independent devices are connected through the operating system to form super devices to achieve resource sharing and functional coordination. However, the management and operation of super devices face many challenges, especially when there are a large number of devices and various types. How to monitor the health status of each sub-device in real time, detect device anomalies in time and handle anomalies has become an urgent problem to be solved.

[0047] In order to solve the above problems, the present invention proposes a super device fault detection method based on multi-dimensional data analysis. By collecting and analyzing multi-dimensional data of each sub-device of the super device (such as CPU utilization, memory usage, network bandwidth, etc.), as well as events reported by the sub-devices to the operating system, real-time monitoring of the health status of the device can be achieved, and AI technology can be used to analyze the fault and put forward corresponding recovery suggestions.

[0048] The super equipment fault detection method based on multi-dimensional data analysis described in the preferred embodiment of the present invention is as follows: Figure 1 As shown, the super equipment fault detection method based on multi-dimensional data analysis includes the following steps:

[0049] Step S10: Obtain historical multi-dimensional data of the super device, and pre-process the historical multi-dimensional data to obtain multi-dimensional indicator data, wherein the historical multi-dimensional data includes historical device basic indicator data, historical key indicator event data, and historical log data.

[0050] like Figure 2 As shown, the present invention is provided with a data acquisition module, a data storage module, a data analysis module, an event trend analysis module, an AI fault analysis and recovery module and a visualization display module.

[0051] Among them, 1. The data acquisition module regularly collects basic resource data such as CPU utilization, memory usage, network bandwidth, etc. of each sub-device through the network interface, and receives events reported by the sub-device; 2. The data storage module stores the collected data in a distributed database to support rapid retrieval and analysis; 3. The data analysis module processes the collected data in real time, calculates the health status indicators of each sub-device, and analyzes events to generate key area indicators. 4. The event trend analysis module regularly analyzes historical event data, identifies common fault trends, and generates detailed fault event reports. 5. The AI ​​fault analysis and recovery module analyzes faults through training models, provides fault recovery suggestions, and continuously optimizes the model to improve the accuracy of fault analysis and recovery suggestions. 6. The visualization display module displays the analysis results to the operation and maintenance personnel through charts and dashboards, provides an intuitive operation and maintenance management interface, and supports real-time monitoring and fault handling.

[0052] Specifically, historical equipment basic indicator data of the super device is obtained, and the historical equipment basic indicator data is cleaned and normalized to obtain equipment basic indicator data; historical key indicator event data of the super device is obtained, and the historical key indicator event data is subjected to a first data classification process and a first data filtering process to obtain key indicator event data; historical log data of the super device is obtained, and the historical log data is subjected to a second data classification process and a second data filtering process to obtain log data; wherein the multi-dimensional indicator data includes the equipment basic indicator data, the key indicator event data and the log data.

[0053] like Figure 3 As shown, the data collection module includes: 1. Equipment basic indicator collection; 2. Event collection; 3. Log collection; 4. Dynamic adjustment of collection frequency; 5. Intelligent data filtering; 6. Distributed data collection architecture; 7. Data encryption transmission.

[0054] Among them, the basic indicators of the equipment are collected: 1. CPU utilization: The system collects the CPU usage of each sub-device of the super device through regular sampling, and monitors the computing resource usage of each sub-device in real time. Through long-term data analysis, it can identify high-load devices and warn of possible performance bottlenecks in advance. 2. Memory usage: The system collects the memory usage of each sub-device, monitors memory allocation and usage trends, and monitors memory usage to detect abnormalities such as memory leaks in a timely manner to ensure stable system operation. 3. Network bandwidth: The system collects network traffic data of each sub-device in real time, including upload and download bandwidth. By monitoring the network bandwidth, it can identify network congestion and optimize network resource allocation. 4. Storage usage: The system collects the storage usage of each sub-device and monitors the occupancy of storage space. By analyzing the storage usage data, it can warn of insufficient storage space in advance and perform optimization and cleanup.

[0055] The event collection process includes: 1. Diversity of event types: The system collects events reported by sub-devices, covering events in all fields, such as system errors, warnings, information, etc. Each event contains detailed event type, timestamp, event description and other information. 2. Key field events: The system focuses on events in the soft bus field, such as key indicator events such as discovery success rate and connection success rate. Through these event data, the communication status and performance of super devices can be fully understood. 3. Event trend analysis: The system conducts long-term trend analysis on the collected event data, identifies the occurrence patterns of common fault events, and helps operation and maintenance personnel prevent and quickly locate problems in advance.

[0056] The log collection process includes: 1. Log system function: The system's log recording system is provided to the system framework, services, and applications to record important information such as user operations and system operating status. By collecting logs, you can fully understand the operating status of the system and applications. 2. User operation log: The system collects and analyzes user operation logs to record user operations on each sub-device. By analyzing these logs, it can help operation and maintenance personnel understand user usage and optimize user experience. 3. System operation status log: The system collects system operation status logs, including important log information such as start, stop, error, and warning. Through these log data, you can fully understand the system's operating status and historical faults, and help operation and maintenance personnel troubleshoot and locate problems.

[0057] The process of dynamically adjusting the collection frequency includes: dynamically adjusting the frequency of data collection according to the operating status of the device and the results of historical data analysis. For example, when the device is in a high-load state, the data collection frequency can be increased to obtain more fine-grained data; when the device is in a low-load state, the collection frequency can be reduced to reduce system overhead.

[0058] The intelligent data filtering process includes: the system can intelligently filter the collected data to remove redundant and invalid data. For example, for stable running equipment, only data at key time points can be collected to reduce the pressure of data storage and processing.

[0059] The distributed data acquisition architecture process includes: the present invention adopts a distributed data acquisition architecture to distribute data acquisition tasks to various sub-devices, thereby reducing the load of the central node and facilitating improving the efficiency and stability of data acquisition.

[0060] The data encryption transmission process includes: In order to ensure data security, the system can use encryption transmission technology during the data collection process to encrypt the collected data to prevent the data from being stolen or tampered with during the transmission process.

[0061] The data storage module includes: multi-dimensional data storage, event storage, log data storage, data storage architecture, large-scale data storage and retrieval.

[0062] Among them, the multi-dimensional data storage process includes: 1. Efficient storage: The system adopts a distributed storage architecture, which can efficiently store the collected multi-dimensional data, including CPU utilization, memory utilization, network bandwidth, storage utilization, etc. Distributed storage can improve the capacity and speed of data storage to meet the storage needs of large-scale data; 2. Data compression and deduplication: In order to optimize the storage space, the system compresses and deduplicates the collected data, reduces the storage of redundant data, and improves storage efficiency.

[0063] The event storage process includes: 1. Event classification storage: The system will classify and store the collected events according to the event type, such as system errors, warnings, and information. Event classification storage can improve the efficiency of data retrieval and facilitate the rapid location and analysis of specific types of events. 2. Event detail storage: The system stores detailed information for each event, including event type, timestamp, event description, etc. Detailed event information helps operation and maintenance personnel to analyze events and locate faults.

[0064] The log data storage process includes: 1. Log hierarchical storage: The system stores logs at different levels, such as system framework logs, service logs, application logs, etc. Hierarchical storage can improve the management and retrieval efficiency of log data. 2. Log indexing: In order to improve the retrieval speed of log data, the system indexes the log data and supports fast retrieval based on multi-dimensional conditions such as time, event type, and device.

[0065] The data storage architecture process includes: 1. Distributed storage architecture: The system adopts a distributed storage architecture to distribute data storage tasks to multiple storage nodes to improve the scalability and reliability of the storage system. The distributed architecture can effectively handle the storage needs of large-scale data and ensure the stability and availability of data storage. 2. Data backup and recovery: In order to ensure the security of data, the system regularly backs up the stored data and supports rapid data recovery. The backup and recovery mechanism can provide effective protection and recovery means when data is lost or damaged.

[0066] The large-scale data storage and retrieval process includes: 1. High-performance database: The system uses a high-performance database management system to support efficient storage and retrieval of large-scale data. High-performance databases can handle a large number of concurrent data requests to ensure the speed and reliability of data storage and retrieval. 2. Distributed file system: In order to further improve the performance of data storage, the system uses a distributed file system to support the storage and management of large-scale data. The distributed file system can provide high-throughput and low-latency data storage services.

[0067] like Figure 4 As shown, the data analysis module includes: real-time analysis of multi-dimensional data, event analysis, and log data analysis.

[0068] The real-time analysis process of multi-dimensional data includes: data preprocessing and health status indicator calculation, where data preprocessing includes: 1. Data cleaning: The system cleans the collected multi-dimensional data, removes noise data and outliers, and ensures the accuracy and consistency of the data. 2. Data normalization: In order to facilitate analysis and calculation, the system normalizes the data and converts data of different dimensions into the same scale range.

[0069] The calculation of health status indicators includes: 1. CPU utilization analysis: The system calculates the CPU utilization of each sub-device in real time, and monitors high-load devices by setting thresholds (such as 80%). The system will mark the devices that exceed the threshold and generate health status indicators. 2. Memory usage analysis: The system analyzes the memory usage of each sub-device, identifies possible memory leaks or insufficient memory by monitoring memory usage (such as 80%), and calculates memory health status indicators. 3. Network bandwidth analysis: The system monitors the network bandwidth usage of sub-devices in real time, identifies network congestion or abnormal traffic by analyzing upload and download bandwidth, and generates network health status indicators. 4. Storage usage analysis: The system calculates the storage usage of each sub-device, monitors the occupancy of storage space by setting thresholds (such as 85%), and calculates storage health status indicators.

[0070] like Figure 4As shown, event analysis includes event data preprocessing and key field indicator calculation. Event data preprocessing includes: 1. Event classification: The system classifies the collected events and organizes them by event type (error, warning, information, etc.). 2. Event filtering: The system filters out low-priority or irrelevant events and retains key events for subsequent analysis.

[0071] The calculation of key area indicators includes: 1. Soft bus discovery success rate: The system analyzes the event data in the soft bus area, calculates the number of successful discovery events and the total number of events, and obtains the discovery success rate indicator. 2. Soft bus connection success rate: The system analyzes the soft bus connection events, calculates the number of successful connection events and the total number of events, and obtains the connection success rate indicator.

[0072] like Figure 4 As shown, log data analysis includes log data preprocessing and operation status analysis, where log data preprocessing includes: 1. Log classification: The system classifies log data and organizes them by log type (system log, application log, user operation log, etc.). 2. Log filtering: The system filters out redundant and irrelevant logs and retains key log data for analysis.

[0073] Operation status analysis includes: 1. System status monitoring: The system analyzes system log data, monitors the system's operation status and historical events, and calculates system health status indicators through log data. 2. User behavior analysis: The system analyzes user operation logs, identifies user operation behaviors and operation modes, and calculates indicators related to user experience.

[0074] Step S20: Obtain multi-dimensional fault data from the multi-dimensional indicator data, and train a preset deep learning model based on the multi-dimensional fault data to obtain a fault analysis model.

[0075] The event trend analysis module includes: event trend analysis, event trend identification, common fault trend identification, and fault cause location.

[0076] Specifically, the basic indicator data of the equipment, the key indicator event data and the log data are aggregated to obtain aggregated multi-dimensional indicator data; and the aggregated multi-dimensional indicator data is converted into a time series to obtain multi-dimensional time series data.

[0077] Event trend analysis includes: data preprocessing and time series analysis. Data preprocessing includes: event classification and aggregation. The system first classifies the collected events, aggregates them according to attributes such as event type, device type, and timestamp, and forms a multi-dimensional event data set. Time series analysis includes: converting the aggregated event data into time series data and arranging them in chronological order to facilitate subsequent trend analysis and pattern recognition.

[0078] Obtain a preset baseline model, and input the multi-dimensional time series data into the preset baseline model, use an anomaly detection algorithm to identify abnormal events in the multi-dimensional time series data through the preset baseline model, and obtain abnormal event data in the multi-dimensional time series data; obtain a preset fault mode library, and perform fault search in the preset fault mode library according to the abnormal event data to obtain multi-dimensional fault data corresponding to the abnormal event data; reprocess the multi-dimensional fault data to obtain a multi-dimensional fault training set, wherein the reprocessing includes data cleaning processing, feature extraction processing and standardization processing; determine a preset deep learning model, and use cross-validation method and hyperparameter tuning method to train the preset deep learning model according to the multi-dimensional fault training set to obtain a fault analysis model. The preset deep learning model includes a convolutional neural network, a recursive neural network and a long short-term memory neural network.

[0079] like Figure 5 As shown, the event trend identification process includes: time window analysis and anomaly detection. Among them, the time window analysis includes: 1. Short-term trend analysis: The system analyzes the event data within a short time window (such as hours and days) to identify the frequency and fluctuation of events in the short term. By calculating the moving average and smoothing curve, the short-term trend of the event is presented; 2. Long-term trend analysis: The system analyzes the event data within a long time window (such as weeks and months) to identify the long-term change trend of the event. By accumulating the number of event occurrences and the rate of change, the long-term trend of the event is presented. Anomaly detection includes: 1. Baseline model construction: The system builds a baseline model based on historical event data to determine the normal event frequency range. The baseline model can be constructed through statistical analysis or machine learning methods; 2. Abnormal event identification: The system compares the current event data with the baseline model to identify abnormal events that exceed the normal range. Abnormal events can be identified by threshold setting or anomaly detection algorithms.

[0080] Common fault trend identification includes: fault pattern identification and trend chart presentation. Among them, fault pattern identification includes: 1. Fault pattern library: The system establishes a common fault pattern library, which contains different types of fault patterns and their characteristics. Fault patterns can be defined through historical data analysis or expert knowledge; 2. Pattern matching: The system matches the current event data with the fault pattern library to identify events that meet specific fault patterns. Through pattern matching, common fault trends can be quickly identified. Trend chart presentation includes: Fault trend chart: The system generates a fault trend chart to show the frequency and change trend of common faults in different time windows. The chart can be filtered and classified by dimensions such as fault type and equipment type.

[0081] Fault cause location includes: event correlation analysis and fault detailed information display. Among them, event correlation analysis includes: 1. Event chain analysis: The system connects related events in series through the event chain analysis method to form an event chain. By analyzing the event chain, the cause and effect of the fault can be identified. 2. Causal relationship analysis: The system uses causal relationship analysis methods (such as Granger: causality test) to identify the causal relationship between events and locate the cause of the fault. Causal relationship analysis can be implemented through statistical methods or machine learning algorithms. Fault detailed information display includes: 1. Event details page: The system provides an event details page to display detailed information of each fault event, including event type, timestamp, device information, event description, etc. Through the event details page, operation and maintenance personnel can have a deep understanding of the specific situation of the fault. 2. Related logs and data: The system associates the relevant logs and data of the fault event, and provides system logs, operation logs, resource usage data, etc. when the fault occurs. By viewing the relevant logs and data, operation and maintenance personnel can further analyze the cause of the fault.

[0082] Step S30: acquiring current multi-dimensional fault data, inputting the current multi-dimensional fault data into the fault analysis model, and outputting the current fault analysis result of the super device.

[0083] like Figure 6 As shown in the figure, the AI ​​fault analysis and recovery module includes: fault analysis combined with AI technology, fault recovery suggestion generation, and historical fault handling experience learning.

[0084] Specifically, current multi-dimensional fault data is acquired, and the current multi-dimensional fault data is input into the fault analysis model; the fault analysis model is used to perform fault analysis on the current multi-dimensional fault data to obtain a fault analysis result, wherein the fault analysis result includes the fault cause and the fault impact range.

[0085] Among them, fault analysis combined with AI technology includes fault data input, deep learning model training and fault cause analysis.

[0086] Fault data input includes: Multi-source data fusion: The system collects multi-source data related to faults, including events, logs, equipment status data, etc., and inputs them into the AI ​​model for analysis. Multi-source data fusion helps provide comprehensive fault information and improves analysis accuracy.

[0087] Deep learning model training includes: 1. Data preprocessing: The system preprocesses the input data, including data cleaning, feature extraction and standardization, to ensure that the data is suitable for the training and reasoning of the deep learning model; 2. Model selection: The system uses deep learning models (such as convolutional neural network CNN, recurrent neural network RNN, long short-term memory network LSTM) for fault analysis. The model is trained with a large amount of historical fault data to learn fault modes and characteristics; 3. Model training: The system uses large-scale historical data to train the deep learning model, optimize the model parameters, and improve the model's fault identification and analysis capabilities. Cross-validation and hyperparameter tuning methods are used during the training process to ensure the generalization ability and accuracy of the model.

[0088] Fault cause analysis includes: 1. Model reasoning: The system inputs real-time fault data into the trained deep learning model and performs fault cause analysis through model reasoning. The model can automatically identify the root cause and impact range of the fault; 2. Explanatory analysis: The system combines explanatory AI technology (such as SHAP, SHapley Additive exPlanation, explainable machine learning; LIME, R language explainable machine learning) to provide explainable analysis of model output to help operation and maintenance personnel understand the cause of the fault and the analysis process.

[0089] Furthermore, a fault recovery suggestion library is obtained, and a search is performed in the fault recovery suggestion library according to the fault cause and the fault impact range to obtain a target fault recovery suggestion.

[0090] Fault recovery suggestion generation includes a recovery suggestion library, an intelligent recommendation system, and suggestion priority sorting. The recovery suggestion library includes: 1. Predefined suggestions: The system establishes a fault recovery suggestion library, which contains predefined recovery suggestions for common faults. Suggestions can include restarting services, adjusting configurations, upgrading software, checking hardware, and other operations. 2. Self-learning suggestions: The system combines machine learning algorithms (such as collaborative filtering, KNN, K-nearst neighbors, nearest neighbor rule classification) to learn from historical fault handling records to generate new recovery suggestions, and continuously update and improve the suggestion library.

[0091] The intelligent recommendation system includes: 1. Real-time recommendation: When a fault occurs, the system generates recovery suggestions in real time and provides personalized recovery suggestions to operation and maintenance personnel through the intelligent recommendation system (such as content-based recommendation and collaborative filtering-based recommendation). 2. Multi-dimensional recommendation: The system combines multi-dimensional information such as fault type, device status, and historical records to generate multiple recovery plans for operation and maintenance personnel to choose and refer to.

[0092] The priority sorting of suggestions includes: 1. Utility evaluation: The system evaluates the utility of each recovery suggestion, and the evaluation indicators may include recovery success rate, time cost, resource consumption, etc. Through the utility evaluation, the system prioritizes the recovery suggestions and recommends the best recovery plan; 2. Feedback mechanism: The system establishes a feedback mechanism to record the adoption and effect of the recovery suggestions by the operation and maintenance personnel, and uses the feedback data for the continuous optimization and improvement of the model.

[0093] Historical fault handling experience learning includes experience accumulation and model optimization. Among them, experience accumulation includes: 1. Fault handling records: The system records the detailed information of each fault handling, including fault type, handling steps, recovery results, handling time, etc. The recorded data is used for subsequent analysis and learning. 2. Experience library construction: The system builds an experience library based on fault handling records to accumulate the handling experience and best practices of operation and maintenance personnel. The experience library can be continuously expanded and updated to form knowledge accumulation.

[0094] Model optimization includes: 1. Continuous learning: The system combines deep learning technology to continuously learn and optimize the model using new fault data and processing records to improve the model's fault analysis and recovery suggestion capabilities; 2. Online update: The system supports online update of model parameters and suggestion libraries to ensure that the model and suggestion library are always kept up to date and optimal. Through online updates, the system can quickly adapt to new fault types and processing requirements.

[0095] The present invention also provides a visualization display module, wherein the visualization display module includes: overall health status visualization, detailed status display of each sub-device, event trend visualization, fault trend visualization, and fault detail visualization.

[0096] Dashboards are set up for overall health status visualization, including: 1. Overview dashboard: Provides an overview of the overall health status of the super device, including real-time monitoring of key indicators such as CPU utilization, memory usage, network bandwidth, and storage usage. Different colors and progress bars are used to indicate the health status of each indicator. 2. Health score: Generate a health score (such as a percentage score) based on various indicators so that operation and maintenance personnel can understand the overall health status of the super device at a glance. The calculation of the health score can be based on weighted average or other statistical methods.

[0097] The detailed status display of each sub-device includes a status chart, which includes: 1. Sub-device status chart: Displays detailed status information of each sub-device, including real-time CPU utilization, memory utilization, network bandwidth, storage utilization, etc. Use line charts, bar charts, etc. to update data in real time; 2. Device status pie chart: Use a pie chart to display the status distribution of sub-devices, such as the number and proportion of normal devices, warning devices, and faulty devices, to help operation and maintenance personnel quickly identify abnormal devices.

[0098] Event trend visualization includes event frequency trend charts, which include: 1. Event line chart: Displays the changes in the frequency of events in different time windows (such as hours, days, and weeks). Uses line charts to display the frequency of different types of events, helping operation and maintenance personnel understand the time distribution and trends of events; 2. Event stacking chart: Uses stacking charts to display the cumulative occurrence of different types of events, making it easier for operation and maintenance personnel to identify the overall trend and fluctuations of events.

[0099] Fault trend visualization includes fault trend charts, which include: 1. Fault bar chart: Displays the frequency of occurrence of different types of faults in each time window, and uses bar charts to compare the number of faults in different time periods to help operation and maintenance personnel identify high-incidence periods of faults; 2. Fault heat map: Uses heat maps to display the distribution of faults in time and equipment dimensions, and uses color depth to indicate the frequency of faults, helping operation and maintenance personnel locate fault hotspots.

[0100] Fault detail visualization includes a fault detail page, which includes: 1. Fault timeline: provides a timeline view of fault events, and displays detailed information of each fault event in chronological order, including event type, occurrence time, affected equipment, fault description, etc., to help operation and maintenance personnel understand the occurrence process of the fault; 2. Fault association diagram: uses a relationship diagram to display the association between fault events, and uses nodes and lines to represent the cause-effect relationship and impact range of different fault events, to help operation and maintenance personnel deeply analyze the cause of the fault.

[0101] In summary, the monitoring device of the present invention has a comprehensive set of functions, aiming to provide a highly integrated and automated management interface for operation and maintenance personnel. The specific functions are as follows:

[0102] 1. Real-time monitoring function: One of the core functions of the system is real-time monitoring, which can continuously track and analyze the operating status of each sub-device of the super device. Through the key indicators such as CPU utilization, memory utilization, network bandwidth and storage utilization collected by the data acquisition module, the system can reflect the basic resource usage of each sub-device in real time. These data are stored through an efficient data storage module and processed in real time through a data analysis module to ensure the accuracy and real-time nature of the monitoring data.

[0103] 2. Event analysis function: The system provides in-depth event analysis functions, which can conduct comprehensive analysis of events and log data. The event trend analysis module can identify the frequency and pattern of events and predict potential failure risks through intelligent algorithms. In addition, the system can classify and filter events, quickly extract key information, and help operation and maintenance personnel understand the context and possible impact of events.

[0104] 3. Fault location and recovery function: Combined with the fault analysis and recovery module of AI technology, the system can quickly locate and deeply analyze faults. Using deep learning models, the system can identify the root cause of the fault and provide customized solutions from the recovery suggestion library. In addition, the system can continuously learn and optimize based on historical fault handling experience to improve the accuracy and efficiency of fault recovery suggestions.

[0105] 4. Visual display function: The system's visual display module provides users with an intuitive operation interface, displaying data and analysis results in the form of charts, dashboards, and timelines. This not only helps operation and maintenance personnel quickly grasp the overall health of the equipment, but also enables them to quickly identify problematic equipment and fault types when faults occur. Through interactive visualization tools, users can drill down to detailed information on events and faults, including the scope of impact, related logs, and recommended recovery operations.

[0106] 5. User experience optimization function: The system also pays special attention to the optimization of user experience. Through user behavior analysis, the system can identify user usage patterns and preferences, thereby providing personalized monitoring views and alarm settings. This allows each operation and maintenance personnel to view and manage the device status in a customized manner according to their own needs and concerns.

[0107] 6. Automated processing function: In order to further improve operation and maintenance efficiency, the system also has an automated processing function, which can automatically trigger a predefined recovery process when a specific type of fault is identified, reduce manual intervention, and speed up fault recovery.

[0108] Technical effects of the present invention:

[0109] 1. Improve operation and maintenance efficiency: The system greatly improves the response speed to equipment status through real-time monitoring and automated data analysis. Operation and maintenance personnel can quickly obtain key indicators of equipment operation, reducing the need for manual inspection, thereby significantly improving operation and maintenance efficiency.

[0110] 2. Enhanced equipment stability: Through in-depth analysis of equipment operation data, the system can predict and identify potential performance bottlenecks and failure risks. This enables the operation and maintenance team to take measures before problems occur, effectively avoiding equipment failures and enhancing equipment stability.

[0111] 3. Ensure equipment reliability: The system combines AI technology to intelligently analyze and diagnose faults and provide accurate fault causes and recovery suggestions. This intelligent fault handling mechanism improves the success rate of fault recovery and ensures equipment reliability.

[0112] 4. Intelligent fault prevention: The system can not only monitor and respond to the current equipment status in real time, but also predict possible problems in the future by learning historical data. This forward-looking fault prevention mechanism further improves the operating efficiency and safety of the equipment.

[0113] 5. Optimize resource allocation: By analyzing the device usage pattern, the system can provide the operation and maintenance team with suggestions for optimizing resource allocation. This helps to allocate computing, storage and network resources more reasonably and improve resource utilization and cost-effectiveness.

[0114] 6. Improve user experience: The system's visual display module provides users with an intuitive and easy-to-use interface, allowing users with non-technical backgrounds to easily understand device status and operation and maintenance activities. This user-centric design improves user experience and satisfaction.

[0115] 7. Support decision-making: The detailed analysis reports and trend forecasts provided by the system provide decision-making support for management. These data-driven insights help to formulate more effective operation and maintenance strategies and business development plans.

[0116] 8. Promote continuous improvement: The system design allows for the continuous integration of new data analysis technologies and AI algorithms, ensuring that the system can adapt to the ever-changing technological environment and business needs. This continuous technological innovation and improvement ensures the long-term competitiveness of the system.

[0117] The technical effects of the present invention are reflected in many aspects. It innovatively combines multi-dimensional data analysis and artificial intelligence technology to provide a comprehensive solution for equipment monitoring and management.

[0118] Furthermore, if Figure 7 As shown, based on the above-mentioned super device fault detection method based on multi-dimensional data analysis, the present invention also provides a super device fault detection system based on multi-dimensional data analysis, wherein the super device fault detection system based on multi-dimensional data analysis includes:

[0119] The data acquisition module 51 is used to acquire the historical multi-dimensional data of the super device and pre-process the historical multi-dimensional data to obtain multi-dimensional indicator data;

[0120] A model training module 52 is used to obtain multi-dimensional fault data in the multi-dimensional indicator data, and train a preset deep learning model according to the multi-dimensional fault data to obtain a fault analysis model;

[0121] The model analysis module 53 is used to obtain current multi-dimensional fault data, input the current multi-dimensional fault data into the fault analysis model, and output the current fault analysis result of the super device.

[0122] Furthermore, if Figure 8 As shown, based on the above-mentioned super device fault detection method and system based on multi-dimensional data analysis, the present invention also provides a terminal accordingly, and the terminal includes a processor 10, a memory 20 and a display 30. Figure 8 Only some components of the terminal are shown, but it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0123] In some embodiments, the memory 20 may be an internal storage unit of the terminal, such as a hard disk or memory of the terminal. In other embodiments, the memory 20 may also be an external storage device of the terminal, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), etc. equipped on the terminal. Further, the memory 20 may also include both an internal storage unit of the terminal and an external storage device. The memory 20 is used to store application software and various types of data installed on the terminal, such as the program code of the installation terminal. The memory 20 may also be used to temporarily store data that has been output or is to be output. In one embodiment, a super device fault detection program 40 based on multi-dimensional data analysis is stored on the memory 20, and the super device fault detection program 40 based on multi-dimensional data analysis can be executed by the processor 10, thereby realizing the super device fault detection method based on multi-dimensional data analysis in the present application.

[0124] In some embodiments, the processor 10 may be a central processing unit (CPU), a microprocessor or other data processing chip, used to run the program code or process data stored in the memory 20, such as executing the super device fault detection method based on multi-dimensional data analysis.

[0125] In some embodiments, the display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, etc. The display 30 is used to display information on the terminal and to display a visual user interface. The components 10-30 of the terminal communicate with each other via a system bus.

[0126] In one embodiment, when the processor 10 executes the super device fault detection program 40 based on multi-dimensional data analysis in the memory 20, the following steps are implemented:

[0127] Acquire historical multi-dimensional data of the super device, and pre-process the historical multi-dimensional data to obtain multi-dimensional indicator data;

[0128] Acquire multi-dimensional fault data from the multi-dimensional indicator data, and train a preset deep learning model according to the multi-dimensional fault data to obtain a fault analysis model;

[0129] Acquire current multi-dimensional fault data, input the current multi-dimensional fault data into the fault analysis model, and output the current fault analysis result of the super device.

[0130] The historical multi-dimensional data includes historical equipment basic indicator data, historical key indicator event data and historical log data;

[0131] The obtaining of historical multi-dimensional data of the super device and preprocessing of the historical multi-dimensional data to obtain multi-dimensional indicator data specifically includes:

[0132] Acquire historical equipment basic indicator data of the super equipment, and perform data cleaning and normalization processing on the historical equipment basic indicator data to obtain equipment basic indicator data;

[0133] Acquire historical key indicator event data of the super device, and perform first data classification processing and first data filtering processing on the historical key indicator event data to obtain key indicator event data;

[0134] Acquire historical log data of the super device, and perform second data classification processing and second data filtering processing on the historical log data to obtain log data;

[0135] Among them, the multi-dimensional indicator data includes the basic indicator data of the equipment, the key indicator event data and the log data.

[0136] The method of obtaining the multi-dimensional fault data in the multi-dimensional indicator data and training a preset deep learning model according to the multi-dimensional fault data to obtain a fault analysis model also includes:

[0137] Aggregate the basic indicator data of the device, the key indicator event data and the log data to obtain aggregated multi-dimensional indicator data;

[0138] The aggregated multi-dimensional indicator data is subjected to time series conversion processing to obtain multi-dimensional time series data.

[0139] The acquiring of the multi-dimensional fault data in the multi-dimensional indicator data and training of a preset deep learning model according to the multi-dimensional fault data to obtain a fault analysis model specifically includes:

[0140] Obtain a preset baseline model, and input the multidimensional time series data into the preset baseline model, and use an anomaly detection algorithm to identify abnormal events on the multidimensional time series data through the preset baseline model to obtain abnormal event data in the multidimensional time series data;

[0141] Acquire a preset fault mode library, and perform fault search in the preset fault mode library according to the abnormal event data to obtain multi-dimensional fault data corresponding to the abnormal event data;

[0142] Reprocessing the fault multi-dimensional data to obtain a fault multi-dimensional training set, wherein the reprocessing includes data cleaning processing, feature extraction processing and standardization processing;

[0143] A preset deep learning model is determined, and a cross-validation method and a hyperparameter tuning method are used to train the preset deep learning model according to the fault multi-dimensional training set to obtain a fault analysis model.

[0144] Among them, the preset deep learning model includes convolutional neural network, recursive neural network and long short-term memory neural network.

[0145] The acquiring of current multi-dimensional fault data, inputting the current multi-dimensional fault data into the fault analysis model, and outputting the current fault analysis result of the super device specifically includes:

[0146] Acquire current multi-dimensional fault data, and input the current multi-dimensional fault data into the fault analysis model;

[0147] The fault analysis model is used to perform fault analysis processing on the current multi-dimensional fault data to obtain a fault analysis result, wherein the fault analysis result includes a fault cause and a fault impact range.

[0148] The method further includes obtaining current multi-dimensional fault data, inputting the current multi-dimensional fault data into the fault analysis model, and outputting the fault analysis result of the super device, and then:

[0149] A fault recovery suggestion library is obtained, and a search is performed in the fault recovery suggestion library according to the fault cause and the fault impact range to obtain a target fault recovery suggestion.

[0150] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a super device fault detection program based on multi-dimensional data analysis, and when the super device fault detection program based on multi-dimensional data analysis is executed by a processor, the steps of the super device fault detection method based on multi-dimensional data analysis as described above are implemented.

[0151] In summary, the present invention provides a super device fault detection method, system, terminal and computer-readable storage medium based on multi-dimensional data analysis, the method comprising: obtaining historical multi-dimensional data of the super device, and pre-processing the historical multi-dimensional data to obtain multi-dimensional indicator data; obtaining multi-dimensional data of faults in the multi-dimensional indicator data, and training a preset deep learning model according to the multi-dimensional data of faults to obtain a fault analysis model; obtaining current multi-dimensional fault data, inputting the current multi-dimensional fault data into the fault analysis model, and outputting the current fault analysis result of the super device. The present invention obtains multi-dimensional data of each sub-device in the super device, and combines AI technology to perform intelligent analysis on the multi-dimensional data, so as to quickly predict and identify abnormal data and abnormal conditions of the equipment in the multi-dimensional data, and provide accurate fault causes and recovery suggestions, which greatly improves the response speed to the equipment status, so that the operation and maintenance personnel can quickly obtain the key indicators of the equipment operation, reduce the need for manual inspection, and thus significantly improve the efficiency of operation and maintenance; at the same time, it also effectively improves the stability of the operation of the super device.

[0152] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or terminal including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or terminal. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the presence of other identical elements in the process, method, article or terminal including the element.

[0153] Of course, those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing related hardware (such as a processor, a controller, etc.) through a computer program, and the program can be stored in a computer-readable storage medium that can be read by a computer, and the program can include the processes of the above-mentioned method embodiments when executed. The computer-readable storage medium can be a memory, a disk, an optical disk, etc.

[0154] It should be understood that the application of the present invention is not limited to the above examples. For ordinary technicians in this field, improvements or changes can be made based on the above description. All these improvements and changes should fall within the scope of protection of the claims attached to the present invention.

Claims

1. A super equipment fault detection method based on multi-dimensional data analysis, characterized in that: The super equipment fault detection method based on multi-dimensional data analysis includes: Acquire historical multi-dimensional data of the super device, and pre-process the historical multi-dimensional data to obtain multi-dimensional indicator data; The historical multi-dimensional data includes historical equipment basic indicator data, historical key indicator event data and historical log data; The obtaining of historical multi-dimensional data of the super device and preprocessing of the historical multi-dimensional data to obtain multi-dimensional indicator data specifically includes: Acquire historical equipment basic indicator data of the super equipment, and perform data cleaning and normalization processing on the historical equipment basic indicator data to obtain equipment basic indicator data; The basic device indicator data includes CPU utilization, memory utilization, network bandwidth and storage utilization; Acquire historical key indicator event data of the super device, and perform first data classification processing and first data filtering processing on the historical key indicator event data to obtain key indicator event data; The key indicator event data includes event type diversity, key field events and event trend analysis; Acquire historical log data of the super device, and perform second data classification processing and second data filtering processing on the historical log data to obtain log data; The log data includes user operation logs and system operation status logs; The multi-dimensional indicator data includes the basic indicator data of the device, the key indicator event data and the log data; Acquire multi-dimensional fault data from the multi-dimensional indicator data, and train a preset deep learning model according to the multi-dimensional fault data to obtain a fault analysis model; The acquiring of the multi-dimensional fault data in the multi-dimensional indicator data, and training of a preset deep learning model according to the multi-dimensional fault data to obtain a fault analysis model, further includes: Aggregate the basic indicator data of the device, the key indicator event data and the log data to obtain aggregated multi-dimensional indicator data; Performing time series conversion processing on the aggregated multi-dimensional indicator data to obtain multi-dimensional time series data; The acquiring of the multi-dimensional fault data in the multi-dimensional indicator data, and training a preset deep learning model according to the multi-dimensional fault data to obtain a fault analysis model specifically includes: Obtain a preset baseline model, and input the multidimensional time series data into the preset baseline model, and use an anomaly detection algorithm to identify abnormal events on the multidimensional time series data through the preset baseline model to obtain abnormal event data in the multidimensional time series data; Acquire a preset fault mode library, and perform fault search in the preset fault mode library according to the abnormal event data to obtain multi-dimensional fault data corresponding to the abnormal event data; The preset failure mode library includes different types of failure modes, and the failure modes are defined by historical data analysis or expert knowledge; Reprocessing the fault multi-dimensional data to obtain a fault multi-dimensional training set, wherein the reprocessing includes data cleaning processing, feature extraction processing and standardization processing; Determine a preset deep learning model, and use a cross-validation method and a hyperparameter tuning method to train the preset deep learning model according to the fault multi-dimensional training set to obtain a fault analysis model; Acquire current multi-dimensional fault data, input the current multi-dimensional fault data into the fault analysis model, and output the current fault analysis result of the super device.

2. The super equipment fault detection method based on multi-dimensional data analysis according to claim 1 is characterized in that: The preset deep learning model includes a convolutional neural network, a recursive neural network, and a long short-term memory neural network.

3. The super equipment fault detection method based on multi-dimensional data analysis according to claim 1 is characterized in that: The obtaining of current multi-dimensional fault data, inputting the current multi-dimensional fault data into the fault analysis model, and outputting the current fault analysis result of the super device specifically includes: Acquire current multi-dimensional fault data, and input the current multi-dimensional fault data into the fault analysis model; The fault analysis model is used to perform fault analysis processing on the current multi-dimensional fault data to obtain a fault analysis result, wherein the fault analysis result includes a fault cause and a fault impact range.

4. The super equipment fault detection method based on multi-dimensional data analysis according to claim 3 is characterized in that: The method further comprises: obtaining current multi-dimensional fault data, inputting the current multi-dimensional fault data into the fault analysis model, and outputting the fault analysis result of the super device, and then further comprising: A fault recovery suggestion library is obtained, and a search is performed in the fault recovery suggestion library according to the fault cause and the fault impact range to obtain a target fault recovery suggestion.

5. A super equipment fault detection system based on multi-dimensional data analysis, characterized in that: The super equipment fault detection system based on multi-dimensional data analysis includes: A data acquisition module is used to acquire historical multi-dimensional data of the super device and pre-process the historical multi-dimensional data to obtain multi-dimensional indicator data; The historical multi-dimensional data includes historical equipment basic indicator data, historical key indicator event data and historical log data; The obtaining of historical multi-dimensional data of the super device and preprocessing of the historical multi-dimensional data to obtain multi-dimensional indicator data specifically includes: Acquire historical equipment basic indicator data of the super equipment, and perform data cleaning and normalization processing on the historical equipment basic indicator data to obtain equipment basic indicator data; The basic device indicator data includes CPU utilization, memory utilization, network bandwidth, and storage utilization; Acquire historical key indicator event data of the super device, and perform first data classification processing and first data filtering processing on the historical key indicator event data to obtain key indicator event data; The key indicator event data includes event type diversity, key field events, and event trend analysis; Acquire historical log data of the super device, and perform second data classification processing and second data filtering processing on the historical log data to obtain log data; The log data includes user operation logs and system operation status logs; The multi-dimensional indicator data includes the basic indicator data of the device, the key indicator event data and the log data; A model training module, used to obtain multi-dimensional fault data in the multi-dimensional indicator data, and train a preset deep learning model according to the multi-dimensional fault data to obtain a fault analysis model; The acquiring of the multi-dimensional fault data in the multi-dimensional indicator data, and training of a preset deep learning model according to the multi-dimensional fault data to obtain a fault analysis model, further includes: Aggregate the basic indicator data of the device, the key indicator event data and the log data to obtain aggregated multi-dimensional indicator data; Performing time series conversion processing on the aggregated multi-dimensional indicator data to obtain multi-dimensional time series data; The acquiring of the multi-dimensional fault data in the multi-dimensional indicator data, and training a preset deep learning model according to the multi-dimensional fault data to obtain a fault analysis model specifically includes: Obtain a preset baseline model, and input the multidimensional time series data into the preset baseline model, and use an anomaly detection algorithm to identify abnormal events on the multidimensional time series data through the preset baseline model to obtain abnormal event data in the multidimensional time series data; Acquire a preset fault mode library, and perform fault search in the preset fault mode library according to the abnormal event data to obtain multi-dimensional fault data corresponding to the abnormal event data; Wherein, the preset fault mode library includes different types of fault modes, and the fault modes are defined by historical data analysis or expert knowledge; Reprocessing the fault multi-dimensional data to obtain a fault multi-dimensional training set, wherein the reprocessing includes data cleaning processing, feature extraction processing and standardization processing; Determine a preset deep learning model, and use a cross-validation method and a hyperparameter tuning method to train the preset deep learning model according to the fault multi-dimensional training set to obtain a fault analysis model; The model analysis module is used to obtain current multi-dimensional fault data, input the current multi-dimensional fault data into the fault analysis model, and output the current fault analysis result of the super device.

6. A terminal, characterized in that: The terminal includes: a memory, a processor, and a super device fault detection program based on multi-dimensional data analysis stored in the memory and executable on the processor. When the super device fault detection program based on multi-dimensional data analysis is executed by the processor, the steps of the super device fault detection method based on multi-dimensional data analysis as described in any one of claims 1 to 4 are implemented.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a super device fault detection program based on multi-dimensional data analysis, and when the super device fault detection program based on multi-dimensional data analysis is executed by a processor, the steps of the super device fault detection method based on multi-dimensional data analysis according to any one of claims 1 to 4 are implemented.

Images