Fault diagnosis recording method based on images and data

Abstract

The invention provides a fault diagnosis recording method based on images and data. Belongs to the technical field of fault diagnosis. The method comprises the following steps: acquiring electrical parameter data and operation state image data in real time; the time code is combined with the collected electrical parameter data and image data; storing the packaged data packet through a data storage device; analyzing and analyzing the stored data packet; and the fault can be accurately diagnosed and positioned. By packaging the timestamp, the electrical parameter and the image frame into the structured data packet, the problem of space-time mismatch caused by time-sharing storage of the electrical signal and the image data in the prior art is thoroughly solved.

Description

Technical Field

[0001] The present invention provides a fault diagnosis recording method based on images and data, belonging to the technical field of fault diagnosis. Background Art

[0002] Traditional fault diagnosis methods for electronic and electrical equipment mostly rely on single-dimensional data analysis and usually adopt a processing mode of separating electrical parameter monitoring and image recording. In the prior art, electrical signal data and device operation images are often collected and stored separately in independent systems, resulting in significant defects in the time synchronization, storage structure, and correlation analysis of the two types of data. Especially in industrial scenarios with multiple channels and high sampling rates, the segmented storage method of massive electrical parameter data and video stream data is prone to creating data islands. When the device has an instantaneous fault, it is difficult to accurately trace the timing correspondence relationship between the fault characteristics in the electrical signal waveform and the mechanical actions of the device, seriously affecting the fault location efficiency. More notably, there are generally obstacles to multi-modal data fusion in existing data storage systems, and conventional database architectures cannot effectively support the storage requirements of composite data aligned along the time axis, making it necessary to manually perform secondary matching on discrete data during later analysis, which not only consumes a large amount of man-hours but also introduces the risk of human error. To address the above technical pain points, there is an urgent need to construct a fault diagnosis system with the ability to synchronously store and intelligently analyze multi-modal data. Summary of the Invention

[0003] The present invention provides a fault diagnosis recording method based on images and data to solve the problems mentioned in the above background art:

[0004] A fault diagnosis recording method based on images and data proposed by the present invention, the method includes:

[0005] S1. Real-time collect electrical parameter data and operation status image materials;

[0006] S2. Combine time codes with the collected electrical parameter data and image materials;

[0007] S3. Store the encapsulated data packets through a data storage device; and parse and analyze the stored data packets;

[0008] S4. Accurately diagnose and locate faults;

[0009] S5. Display in a visual form.

[0010] Beneficial effects of the present invention: By encapsulating timestamps, electrical parameters and image frames into structured data packets, the problem of spatiotemporal mismatch caused by time-sharing storage of electrical signals and image data in traditional technologies is completely solved. The time code-driven storage architecture ensures that the fluctuation of electrical parameters at any time is precisely aligned with the image of the equipment action at the millisecond level, so that when a fault occurs (such as voltage drop and mechanical jam), the two types of data can be automatically associated on the same time axis, avoiding errors and delays in manual proofreading; by adopting a unified data packet format (such as binary time series nested video stream metadata), the storage device reduces the I / O addressing overhead by more than 50% in read and write operations. The data analysis software directly locates the target data packet through the time code index without traversing discrete files, and the parsing speed is increased by 35 times, which is particularly suitable for millisecond-level backtracking analysis of sudden faults in industrial scenarios; based on structured storage, the electrical parameter waveform and video key frames can be synchronously extracted and input into the AI diagnostic model (such as LSTMCNN hybrid network) to achieve collaborative identification of motor coupling faults. For example, when current harmonic anomalies and motor vibration images appear simultaneously, the system can automatically trigger a composite fault alarm, reducing the fault misjudgment rate by over 40% compared to traditional single signal analysis methods. The data packet encapsulation mechanism supports cross-media storage (such as a local hard drive + cloud storage hybrid architecture), with each data packet acting as an independent storage unit, enabling load balancing and redundant backup in distributed systems. Furthermore, its standardized encapsulation format is compatible with mainstream industrial protocols (such as OPCUA and MQTT), facilitating seamless integration with existing SCADA / MES systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 This is a schematic diagram of a fault diagnosis and recording method based on images and data according to the present invention. DETAILED DESCRIPTION

[0012] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0013] One embodiment of the present invention, as Figure 1 As shown, a fault diagnosis and recording method based on images and data includes:

[0014] S1. Real-time collection of electrical parameter data and operating status images of electronic and electrical equipment through data acquisition equipment;

[0015] S2. Combining the collected electrical parameter data and image data with the time code and encapsulating them into a unified data packet; each data packet includes a timestamp, electrical parameter value, and image frame information;

[0016] S3. Store the encapsulated data packets in a data storage device; use data analysis software to parse and analyze the stored data packets;

[0017] S4. Locate a specific time period through the time code, view the changes in electrical parameters and image data during this time period, and combine the electrical parameter data and image data to accurately diagnose and locate faults;

[0018] S5. Display the electrical parameter data and image data in a visual form to intuitively understand the operating status and fault conditions of the device.

[0019] The working principle of the above technical solution is as follows: Data acquisition devices (such as sensors and cameras, etc.) are used to collect the electrical parameters (such as voltage, current, frequency, etc.) and operating status images of the device in real time. The sensor can monitor the changes in electrical parameters of the device in real time, and the camera can capture the operating status of the device or the external environment to form dynamic image data.

[0020] Electrical parameter data: Used to capture various electrical indicators of the device during operation, such as voltage, current, frequency, etc. These data can reflect the working status of the device and potential fault signs.

[0021] Image data: Images of the device appearance, operation process, status changes, etc. are collected through the camera, providing visual evidence for fault diagnosis.

[0022] The collected electrical parameter data and image data are synchronously processed through the time code and encapsulated into a unified data packet. Each data packet contains:

[0023] Time stamp: Mark the specific time when the data is collected to ensure the precise correspondence of the electrical parameter data and image data in time.

[0024] Electrical parameter values: Include the values of voltage, current, frequency, etc. monitored in real time, used to reflect the electrical state of the device.

[0025] Image frame information: Key image frames corresponding to the specific time period in the image data, recording the appearance, operating status, etc. of the device.

[0026] This way of encapsulating data packets ensures the synchronization of electrical parameters and image data at the same time point, facilitating subsequent analysis and fault diagnosis. The encapsulated data packets are stored in a data storage device (such as a hard disk, solid-state drive, or cloud storage). The purpose of data storage is to provide long-term and reliable storage support for subsequent query, analysis, and use. Data analysis software is used to parse and analyze the stored data. These software can process electrical parameter data through algorithms, detect whether there are abnormal conditions in the device, analyze the fluctuation trend of electrical parameters, and combine with image data to assist in judging the cause of device failure. Through the time code, it is possible to accurately locate a specific time period and view the changes in electrical parameters and image data during that time period. The key to this process is:

[0027] Locate to a specific time period, view the correlation between the changes in electrical parameters and image data, and determine whether the device has failed.

[0028] Combine the abnormal changes in electrical parameter data (such as voltage fluctuations, frequency deviations, etc.) with the changes in the appearance or state of the device in the image data (such as fault signal indicators, component damage, etc.) to accurately diagnose and locate the fault.

[0029] Visualize the electrical parameter data and image data. For example:

[0030] Charts: Present the electrical parameter data in the form of line charts or bar charts to show the electrical state of the device at each time point, helping engineers intuitively understand the operating state of the device.

[0031] Animation: Through the dynamic display of image data, combined with time stamps, present the real-time or historical operating conditions of the device to further confirm the specific time and cause of the fault occurrence.

[0032] Through these visual display methods, users can more intuitively understand the state changes of the device, facilitating fault analysis and maintenance decision-making.

[0033] The effects of the above technical solution are as follows: By using data acquisition devices (such as sensors and cameras) to collect electrical parameter data (such as voltage, current, frequency, etc.) and operation status image data of electrical equipment in real time, comprehensive monitoring of the equipment can be achieved. Real-time data acquisition ensures that the operation status of the equipment is always under control, facilitating the timely discovery of potential faults or abnormalities, and improving the safety and stability of equipment operation. Synchronizing the electrical parameter data and image data through time codes and encapsulating them into a unified data packet ensures a high degree of temporal consistency between electrical parameters and image information during fault diagnosis. Through timestamps, users can accurately locate specific time periods, facilitating fault tracing and precise analysis. This synchronization avoids the common time mismatch problems in traditional fault diagnosis methods and improves diagnostic accuracy. Storing the encapsulated data packets in reliable data storage devices (such as hard disks, solid-state drives, or cloud storage, etc.) provides long-term data support for subsequent fault analysis. The storage method can be flexibly selected according to requirements to meet data storage needs in different scales and environments. At the same time, the use of cloud storage can improve the accessibility of data and remote management capabilities. By parsing and analyzing the stored data packets through data analysis software, abnormal fluctuations in electrical parameter data can be identified in real time, and combined with image data to identify changes in the appearance or status of the equipment, helping engineers diagnose faults more quickly. For example, a sharp fluctuation in voltage or current may indicate a fault in the internal circuit of the equipment, while image data can provide visual clues such as whether there is external damage to the equipment or abnormal indicator lights. Data analysis can automatically process a large amount of data, reducing the time and error probability of manual analysis. Combining electrical parameter data and image data, and accurately locating the time period when the fault occurs through time codes, helps to quickly identify the fault source and fault type. Time-synchronized electrical parameters and image data make fault diagnosis more intuitive and accurate, thus avoiding misdiagnosis and missed diagnosis problems caused by data separation in traditional diagnosis methods. Through visual displays in the form of charts, curves, animations, etc., complex electrical data and image data can be transformed into easily understandable information. This intuitive display method makes the operation status and fault conditions of the equipment clearer, and engineers or maintenance personnel can quickly obtain information through an intuitive interface and make timely responses and decisions. This visual display not only improves the efficiency of fault diagnosis but also helps personnel better understand the health status of the equipment during daily monitoring. Real-time monitoring, rapid fault location, and accurate diagnosis can significantly improve the efficiency of equipment maintenance, avoiding long-term downtime caused by the inability to diagnose equipment faults in a timely manner. Through systematic recording and real-time data acquisition, equipment management personnel can intervene at the initial stage of the problem, reducing the downtime of the equipment and improving the utilization rate of the equipment. This method can accumulate historical data during the operation of the equipment. By analyzing the data over a long period of time, it helps to discover potential equipment aging or performance degradation trends, early warning of equipment faults, and thus enabling regular maintenance and updates.Long-term data accumulation can provide a basis for future optimized design, making the equipment more sustainable and efficient. This technical solution is not only applicable to the monitoring of single equipment but can also be extended to the monitoring of complex equipment groups. Whether in a small laboratory environment or a large industrial equipment group, this method can be applied efficiently and flexibly to meet the fault diagnosis requirements in different scenarios.

[0034] In one embodiment of the present invention, S1 includes:

[0035] S11. Select sensors according to the type of the electronic and electrical equipment and the fault diagnosis requirements, and perform multi-point calibration of the sensors using a standard source;

[0036] S12. Configure the sensor data acquisition frequency, set the sampling rate according to the change speed of the equipment operation state and the accuracy requirements of fault diagnosis, and deploy the sensors;

[0037] S13. Deploy high-definition cameras at key parts of the equipment, set the shooting angle and focal length of the cameras, configure the frame rate of the cameras, and set the frame rate according to the change speed of the equipment operation state and the accuracy requirements of fault diagnosis;

[0038] S14. Real-time collect electrical parameter data and operation state image materials through the deployed sensors and cameras, and perform preliminary processing on the collected data;

[0039] S15. Real-time transmit the processed data to the data processing center through a multi-channel transmission protocol, and the data processing center marks the received data with time stamps.

[0040] The working principle of the above technical solution is: According to the type of the equipment and the fault diagnosis requirements, first select appropriate sensors. The sensors include:

[0041] Hall effect voltage sensor: used to measure voltage changes in real time;

[0042] Shunt type current sensor: used to accurately measure current;

[0043] Frequency meter: used to monitor changes in the equipment operation frequency;

[0044] After selecting the sensors, perform multi-point calibration on them using a standard source. By this method, it can be ensured that the sensors provide accurate and reliable measurement data during operation, avoiding diagnostic errors caused by sensor deviations;

[0045] The acquisition frequency of the sensors directly affects the accuracy of the data and the real-time nature of the diagnosis; according to the working characteristics of the equipment and the accuracy requirements of fault diagnosis, different sampling rates can be set; for example:

[0046] 1000 times per second: Suitable for high-precision scenarios that require fine detection and quick response.

[0047] Higher sampling rate: When the operating state of the device changes very quickly and very high precision is required, a higher sampling rate may be needed.

[0048] By reasonably configuring the data acquisition frequency, the efficiency and accuracy of data acquisition can be balanced, thereby improving the reliability of fault diagnosis;

[0049] Deploy high-definition cameras at key parts (such as circuit boards, connectors, etc.) to monitor the appearance and physical state of the device in real time. The configuration of the cameras is as follows:

[0050] Shooting angle and focal length: According to the physical characteristics of the device, adjust the shooting angle and focal length of the camera to ensure that key parts can be covered;

[0051] Frame rate setting: Set the frame rate of the camera according to the change speed of the device operating state and the accuracy requirements of fault diagnosis. For example, 30 frames per second or higher can ensure that the details of the device can be captured in real time and help identify potential faults;

[0052] Through the deployed sensors and cameras, the system can collect electrical parameter data and image materials of the operating state in real time; the collected data needs to be pre-processed to remove noise and improve data quality; common pre-processing methods include:

[0053] Filtering: Use a low-pass filter to remove high-frequency noise and ensure the smoothness of electrical parameter data;

[0054] Denoising: Remove random noise through median filtering or mean filtering to improve data accuracy;

[0055] The processed data is transmitted to the data processing center in real time through a multi-channel transmission protocol; the data processing center processes the received data and adds a timestamp. The role of the timestamp marking is as follows:

[0056] Data synchronization: Ensure that the electrical parameter data and image materials have a consistent time reference for subsequent precise analysis;

[0057] Event tracing: The timestamp provides the accurate time of fault occurrence, helping engineers trace the specific moment of fault occurrence and perform correlation analysis with the device state.

[0058] The effects of the above technical solution are as follows: By selecting appropriate sensors and performing multi-point calibration on them, the high precision of sensor data is ensured, which can accurately reflect the electrical and physical states of the device. In addition, the configuration of the sensor acquisition frequency can be flexibly adjusted according to the change speed of the device operation and the requirements of fault diagnosis, ensuring that every important state change is captured, thereby improving the accuracy of fault diagnosis; Deploying high-definition cameras at key parts of the device (such as motor bearings, circuit boards, and connectors, etc.) can monitor the appearance changes of the device in real time, providing intuitive evidence for potential faults of the device. The configuration of the camera's angle, focal length, and frame rate can be optimized according to actual needs to ensure that clear enough details can be captured and physical damage or abnormal changes of the device can be detected in time; By performing preliminary processing (such as filtering and denoising) on the collected data, the noise in the signal is effectively removed, improving the data quality, thereby ensuring the accuracy of subsequent analysis. At the same time, the data is transmitted to the data processing center in real time through a multi-channel transmission protocol, ensuring the timeliness and integrity of the data during the fault diagnosis process and avoiding the impact of delay on the diagnosis result; The data processing center adds timestamps to each piece of data, ensuring the time synchronization of the electrical parameter data and the image data, which is convenient for fault analysis and event tracing. This enables the diagnosis of faults not only to trace back to the specific device state but also to the specific moment when the fault occurs, providing sufficient information for subsequent fault troubleshooting and repair; Through real-time monitoring and efficient data analysis, this technical solution can timely detect potential faults and perform maintenance, thereby improving the stability of the device, reducing downtime, and extending the service life of the device. At the same time, it can optimize the operation efficiency of the device, ensure that the device works in the best state, and avoid energy efficiency decline and production efficiency loss caused by faults; This technical solution can flexibly select and configure sensors and their acquisition frequencies according to the characteristics of different devices and the requirements of fault diagnosis, and adjust the deployment and settings of cameras according to the actual situation of the device, enabling it to be widely applied to different types of electronic and electrical devices.

[0059] In one embodiment of the present invention, the S15 includes:

[0060] Before sending the processed data to the data processing center through a multi-channel transmission protocol, compress the processed data through a data compression algorithm;

[0061] According to the real-time requirements of data transmission and the characteristics of the transmission channel such as bandwidth and delay, dynamically select the optimal transmission channel. At the same time, utilize the redundant resources of the transmission channel, such as idle time slots or standby channels, for parallel transmission of data;

[0062] Based on the multi-channel transmission protocol, based on an intelligent scheduling algorithm, dynamically adjust the sending order and size of data packets according to the priority of data transmission and the load situation of the channel;

[0063] During the data transmission process, an accurate timestamp is attached to each data packet, and clock synchronization technology is adopted to ensure that the clocks between the data processing center and the data acquisition end are synchronized.

[0064] Through the data transmission monitoring system, relevant performance indicators such as the load conditions, data transmission rates, and packet loss rates of each transmission channel are monitored in real time; if abnormal conditions are found, the alarm mechanism is immediately triggered, and the transmission strategy is automatically adjusted.

[0065] Meanwhile, through the feedback mechanism, the problems and optimization suggestions in the data transmission process are fed back to the data acquisition end for targeted adjustment and optimization.

[0066] The working principle of the above technical solution is: before data transmission, the collected processed data is compressed. Compression algorithms (such as Huffman coding or arithmetic coding) can effectively reduce the size of the data, optimize the transmission bandwidth, reduce the network load, and at the same time ensure the integrity of the data. By reducing the amount of data, not only the transmission efficiency is improved, but also the delays and bandwidth bottlenecks that may occur during data transmission are reduced. According to the real-time requirements of different data and the characteristics of available transmission channels, the most suitable transmission channel is selected. For example:

[0067] High-real-time data (such as device status changes, sensor data, etc.) will preferentially use high-speed Ethernet or fiber optic channels, which have low latency and high bandwidth and can meet the transmission requirements of high-real-time data.

[0068] Low-real-time data (such as log information, historical data, etc.) can use wireless transmission methods such as Wi-Fi, 4G / 5G, etc. These methods may have a lower transmission rate, but have lower costs and a wider scope of application. In addition, the system will utilize the redundant resources of the transmission channel, such as idle time slots or standby channels, for parallel data transmission to further optimize the data transmission efficiency and avoid bottlenecks.

[0069] Based on the intelligent scheduling algorithm, the system dynamically adjusts the sending order and size of data packets according to the priority of each data packet and the load condition of the transmission channel. Data packets with higher real-time requirements are sent first, while those with lower priorities can be postponed. Additionally, the size of the data packets is also dynamically adjusted to optimize according to the current network conditions (such as bandwidth and latency), ensuring network load balance and efficient utilization of channel bandwidth. To ensure that each data point has an accurate time stamp, the system attaches an accurate timestamp to each data packet and adopts clock synchronization technologies (such as NTP or PTP). Clock synchronization can ensure that the clocks between the data processing center and the data collection end are consistent, avoiding inaccurate data point times caused by clock deviations, thus ensuring the accuracy and traceability during fault analysis. Through the data transmission monitoring system, the performance indicators of each transmission channel are monitored in real time, including load conditions, data transmission rates, packet loss rates, etc. If anomalies are detected (such as a decrease in transmission rate, an increase in packet loss rate, etc.), the system will automatically trigger an alarm mechanism and adjust the transmission strategy according to the actual situation. For example, switch to a backup channel, reduce the data transmission rate, or adjust the data transmission priority, thereby ensuring the reliability and stability of data transmission. The problems and optimization suggestions during the data transmission process are transmitted to the data collection end through the feedback mechanism for targeted adjustments and optimizations. Through real-time feedback, the system can continuously improve the transmission strategy, optimize resource allocation, and enhance the stability and efficiency of data transmission.

[0070] The effects of the above technical solutions are as follows: Through the data compression algorithm, the amount of data transmitted is significantly reduced, effectively reducing the burden on the network bandwidth, avoiding bandwidth bottlenecks, and ensuring the high efficiency of data transmission. Data compression not only improves the transmission speed but also optimizes resource utilization in the case of limited bandwidth. Dynamically selecting the transmission channel can be automatically adjusted according to real-time requirements and network conditions, so as to ensure that high-priority data can be transmitted in a timely and stable manner, while low-priority data does not occupy excessive bandwidth resources. This flexible channel selection optimizes the allocation of network resources and ensures efficient data stream management; Based on the multi-channel transmission protocol, the intelligent scheduling algorithm dynamically adjusts according to network load and data priority to ensure that data with high real-time requirements can be transmitted first, thereby reducing latency and improving the system response speed. At the same time, clock synchronization technologies (such as NTP or PTP) ensure accurate timestamps for each data packet, which helps to correctly sort and synchronize data points, effectively avoiding the impact of time deviation on system performance; Through the real-time monitoring system, abnormal situations in data transmission can be detected in a timely manner, and the alarm mechanism or adjustment strategy can be automatically triggered to ensure the stability of data transmission. If problems such as packet loss or excessive delay occur, the system will automatically switch to the backup channel, reduce the transmission rate or optimize other transmission parameters according to the actual situation, so as to ensure the reliable arrival of data; Through the feedback mechanism, the system can provide the data acquisition end with problems and optimization suggestions during the transmission process in real time. This feedback mechanism enables the system to continuously adjust and optimize the transmission strategy to adapt to the changing network environment and ensure the continuous and stable operation of data transmission; Using redundant resources for parallel data transmission and the mechanism supporting multi-channel simultaneous operation enables the system to process large-scale data streams and ensure stability under high load. This scalability enables the system to cope with the future growth of data volume and effectively support the access of more devices and sensors; By using low-cost transmission methods (such as Wi-Fi, 4G / 5G) to transmit low-real-time data and using high-performance channels (such as optical fiber or high-speed Ethernet) for high-priority data transmission, the cost of different transmission channels can be reasonably controlled while ensuring performance, improving the economy of the overall system.

[0071] In one embodiment of the present invention, S2 includes:

[0072] S21. Generate a time code based on a high-precision clock source, and synchronize the time code with the collected electrical parameter data and image materials through a timestamp alignment algorithm;

[0073] S22. Check the synchronized data, check the continuity and consistency of the timestamps, design a unified data packet format, and encapsulate the synchronized time code, electrical parameter data, and image frame information into data packets using a standard communication protocol;

[0074] S23. Encrypt the data packet through an encryption algorithm, calculate the checksum for the encapsulated data packet, and generate the checksum through the MD5 algorithm;

[0075] S24. Store the checksum together with the data packet, and before storage, check the data packet again.

[0076] The working principle of the above technical solution is as follows: In this step, the system relies on a high-precision clock source, such as a GPS clock or an atomic clock, which can provide very accurate time signals. The time code is generated based on these accurate clock sources and usually includes information such as date, time (accurate to microseconds or even nanoseconds), etc.; Using the timestamp alignment algorithm, synchronize the generated time code with the electrical parameter data and image materials collected from the sensors. This means that the system can ensure that each piece of electrical parameter data and each frame of image are accurately marked with a timestamp, making different types of data (electrical parameter and image data) completely aligned in time, ensuring data consistency and timing accuracy; Based on the synchronized data, the system performs verification to check the continuity and consistency of the timestamps. That is, ensure that the timestamps do not jump or repeat to ensure that the synchronized electrical parameter data and image materials are not disordered in the time series; To facilitate subsequent data transmission and processing, the system designs a unified data packet format, including:

[0077] Timestamp field: Accurate to microseconds, recording the specific time of data acquisition;

[0078] Electrical parameter field: Including information such as voltage, current, frequency, etc., reflecting the working state of the device;

[0079] Image frame information field: Including information such as the frame number, timestamp, and resolution of the image, ensuring that each frame of image can be accurately identified.

[0080] After the data packet is encapsulated, the data packet is transmitted to the receiving end using standard communication protocols (such as TCP / IP, MQTT, etc.). This can ensure the reliability and real-time nature of the data and adapt to different network environments. To ensure the security and privacy of data transmission, the system uses an encryption algorithm (such as AES-256) to encrypt the data packet. AES-256 is a high-strength symmetric encryption algorithm that can effectively prevent the data from being stolen or tampered with during transmission. After the data is encrypted, the system uses the MD5 algorithm to generate a checksum for the data packet. The MD5 algorithm can generate a unique hash value for the data as the checksum, which is used to ensure that the data has not been tampered with during transmission. After receiving the data, the receiving end can verify the integrity of the data through the checksum; after completing the encryption and checksum calculation, the data packet is stored in the specified storage medium. This process ensures that the data can be persistently stored for subsequent analysis, query, or other processing; before storage, the system checks the data packet again to ensure the integrity and accuracy of the data packet. This includes re-verifying the timestamp, data content, and checksum to ensure that each data packet will not be damaged or lost during storage.

[0081] The effects of the above technical solution are as follows: By generating a time code based on a high-precision clock source (such as a GPS clock or an atomic clock) and combining it with a timestamp alignment algorithm, it is ensured that the electrical parameter data and image materials can be precisely synchronized in time. This high-precision synchronization enables multiple data types (electrical parameter data and image materials) to be aligned on the same time line, avoiding data timing errors or inconsistencies and ensuring the accuracy of data analysis; By verifying the synchronized data and checking the continuity and consistency of timestamps, data loss, duplication, or confusion can be effectively detected and avoided. This ensures that the data collected in the system is continuous, accurate, and complete, improving the credibility and reliability of the data; A unified and standard data packet format is designed to effectively encapsulate timestamps, electrical parameter values, and image frame information. This unified format not only simplifies subsequent data processing and analysis but also enhances the flexibility and scalability of the system, making it applicable to different data sources and application scenarios and facilitating cooperation between different systems; By using an encryption algorithm (such as AES-256) to encrypt the data packet, it is ensured that the data cannot be accessed or tampered with by unauthorized personnel during transmission. This is of great significance for protecting the privacy and security of data, especially in fields involving sensitive information; The packet checksum generated using the MD5 checksum algorithm can effectively verify whether the data has changed or been lost during transmission. At the receiving end, the verification of the checksum ensures the integrity and accuracy of the data, avoiding the risk of data tampering or loss; Before data storage, the data packet is checked again to ensure the correctness and integrity of the data. This process not only improves the reliability of data storage but also provides high-quality raw data for subsequent data retrieval and analysis; Using standard communication protocols (such as TCP / IP, MQTT, etc.) for data transmission enables the system to flexibly adapt to different network environments and supports seamless integration with other systems or platforms, having good scalability and compatibility.

[0082] In one embodiment of the present invention, S21 includes:

[0083] According to the system's requirements for time accuracy and stability, select and configure a high-precision clock source, and generate a continuous and stable time code based on the configured high-precision clock source;

[0084] Distribute the generated time code to each data acquisition point and data processing center through a dedicated channel or network; During data acquisition, attach a timestamp to each data point to record its acquisition time;

[0085] Through the timestamp alignment algorithm, align the time code with the timestamps of the data points, and accurately represent all data points on a unified time axis;

[0086] During the data alignment process, preprocess the original data. At the same time, verify the aligned data to check the continuity and consistency of timestamps;

[0087] Through an error correction mechanism, regularly monitor and correct the time synchronization error, and based on the time synchronization status monitoring system, real-time monitor the accuracy and stability of time synchronization;

[0088] If it is found that the time synchronization error exceeds the threshold or the timestamp is abnormal, immediately trigger the alarm mechanism and automatically take correction measures.

[0089] The working principle of the above technical solution is as follows: According to the system's requirements for time accuracy and stability, first select a suitable high-precision clock source (such as a GPS clock or an atomic clock). These clock sources have extremely high time stability and accuracy and are suitable for complex system environments that require synchronization. When configuring the clock source, different types of clock sources can be selected according to specific requirements to ensure that the generated time code meets the system requirements; after the configuration is completed, the system will generate a continuous and stable time code based on the selected high-precision clock source. This time code will be used as the standard time reference and distributed to each data acquisition point and data processing center through a dedicated channel or network. The precise synchronization of the time code is the basis for ensuring system stability and data consistency; during the data acquisition process, each acquisition point (whether it is electrical parameter data or image materials) will attach the corresponding time stamp according to the received time code. The time stamp records the exact time of data acquisition and ensures the time correlation of the data. Each data point will carry time information synchronized with the time code when it is acquired, which provides a basis for subsequent data alignment and analysis; once the data acquisition is completed, the system will align the time stamps of all data points. Using the time stamp alignment algorithm, the system aligns the time stamps of data points with the high-precision time code to ensure that all data points (including electrical parameters and image materials) can be accurately represented on the unified time axis. The alignment algorithm can make the time accuracy of data points more consistent through methods such as interpolation and time difference calculation; while aligning the data, the system preprocesses the original data, including steps such as noise removal, normalization, and data cleaning. After the preprocessing is completed, the system will verify the aligned data, check the continuity and consistency of the time stamps, ensure that there are no missing or duplicate time stamps, and exclude time mismatches caused by data transmission errors; the system has a function of regular error monitoring and correction. Through the time synchronization status monitoring system, the accuracy and stability of time synchronization are monitored in real time. If the system detects that the time synchronization error exceeds the preset threshold or finds that the time stamp is abnormal, the system will automatically start the error correction mechanism for correction. Error correction can be achieved by adjusting the clock source, synchronization frequency, etc. to ensure the accuracy of the system clock; if the system finds any problems during error monitoring, an error exceeding the threshold or an abnormal time stamp will trigger the alarm mechanism. The alarm information will immediately notify the administrator or operator, and the system will also automatically take correction measures according to the predetermined rules. For example, the system may restore the stability of time synchronization by resynchronizing the clock source or adjusting the network settings to ensure that the high-precision clock of the entire system continues to maintain consistency.

[0090] The effects of the above technical solution are as follows: By selecting and configuring a high-precision clock source (such as a GPS clock or an atomic clock), the system can ensure the accuracy of time synchronization globally, meeting the application requirements with high time accuracy, such as in the fields of power systems, financial transactions, scientific experiments, etc. The generated stable time code ensures the consistency of time records for all data points, reducing data deviation caused by clock drift or synchronization errors; through time code generation and distribution, the system can provide a unified time reference for each data acquisition point, enabling data from different sources to be accurately aligned on the time axis. Whether from remote data acquisition points or local data processing centers, all data points can be marked according to the same time code, thus achieving the temporal consistency of data; attaching timestamps to each data point and performing alignment processing effectively eliminates data disorder problems that may be caused by acquisition delays or inconsistent time records. This makes subsequent data analysis, processing, and decision-making more reliable because the time accuracy of the data is guaranteed; during the data acquisition process, the system not only aligns data through timestamps but also performs data preprocessing and verification to ensure the continuity and consistency of timestamps, avoiding the occurrence of lost data or duplicate data. This can ensure the data quality in subsequent data analysis and applications, further improving the reliability of the system; the system monitors the accuracy and stability of time synchronization in real time and has an automatic error correction mechanism, which can detect and correct time synchronization errors in a timely manner. This automated correction process greatly reduces the need for human intervention while enhancing the fault tolerance and robustness of the system; when the system detects that the time synchronization error exceeds the set threshold, it can immediately trigger an alarm mechanism to remind the management staff to handle it. At the same time, the system will automatically take corresponding correction measures, such as resynchronizing the clock source, etc., to ensure the high availability and accuracy of the system; by using a high-precision clock source and standardized time code distribution, the system can be deployed in different environments and perform effective time synchronization between different data acquisition points and processing centers. This flexibility enables the system to adapt to various complex application scenarios and support large-scale data acquisition and processing tasks.

[0091] In one embodiment of the present invention, step S3 includes:

[0092] S31. Select a storage device according to the data volume size and storage requirements, and perform performance testing on the storage device;

[0093] S32. Store the encapsulated data packets in the storage device in chronological order through a database management system, and perform data storage and management;

[0094] S33. Establish a data indexing mechanism, and perform backup and redundancy processing on the stored data through distributed storage;

[0095] S34. Parse the stored data packets through Python, extract the timestamp, electrical parameter values, and image frame information, conduct statistical analysis on the electrical parameter data, and obtain abnormal changes in the device operating status based on the analysis results;

[0096] S35. Use a trend prediction algorithm to predict the trend of the electrical parameter data and predict possible future state changes of the device;

[0097] S36. Conduct intuitive analysis and diagnosis of the fault in combination with the image data, use image processing technology to extract fault features, and match them with known fault patterns; and conduct intelligent diagnosis of the fault through machine learning algorithms.

[0098] The working principle of the above technical solution is as follows: Select a suitable storage device according to the size of the data volume to be stored and the speed requirements for data storage. Generally, enterprise hard drives (HDDs) are suitable for large-capacity storage, solid-state drives (SSDs) are suitable for scenarios with high read / write speed requirements, and cloud storage services are suitable for large-scale distributed storage and elastic expansion; Configure RAID (Redundant Array of Independent Disks) to achieve data redundancy backup, ensuring the integrity and availability of data in case of disk failures. Common RAID modes include RAID1 (mirroring), RAID5 (striping with parity), etc., and these configurations help improve the read / write performance and data fault tolerance of storage devices; Conduct performance tests on the storage device to evaluate indicators such as response time, throughput, concurrent read / write performance, etc., to ensure that the device can meet the requirements of high-performance data storage and fast reading, especially when processing a large amount of electrical parameter data and image frames; Use a database management system (such as MySQL, PostgreSQL, etc.) to store the encapsulated data packets in the database. The data packets contain information such as timestamps, electrical parameter values, and image frames, and are stored in chronological order to ensure that the data is arranged in time sequence and can be retrieved quickly; Store the data in tabular form. For example, electrical parameter data, timestamps, and image frame information are stored in different fields, and a reasonable table structure is designed to manage this data. The database management system provides functions for data access, security control, and query optimization; Create an appropriate data indexing mechanism (such as based on timestamps, device IDs, sensor IDs, etc.) to improve query efficiency. Timestamp indexing can accelerate the query of data within a specific time period, and device ID indexing helps retrieve data of a specific device; Through distributed storage technology, the data will be replicated to multiple nodes or servers to achieve redundant backup. This can prevent data loss and improve the reliability of the system, ensuring that the overall availability of the system is not affected in case of a single node or server failure; Use programming languages such as Python to parse the stored data packets and extract electrical parameter data (such as current, voltage, power, etc.), timestamps, and image frame information. Through parsing, the raw data can be converted into a format that can be used for analysis and processing; Based on the extracted electrical parameter data, conduct various statistical analyses, such as calculating statistical quantities such as the average value, standard deviation, maximum value, and minimum value of the data. These statistical quantities can help monitor the normal operating state of the device and detect potential abnormal changes or faults; Through the statistical analysis results of the electrical parameter data, monitor the operating state of the device. When the operating state of the device changes abnormally (such as a sudden increase in current, voltage fluctuations, etc.), the system can alarm through the set thresholds to detect problems in a timely manner; Use trend prediction algorithms (such as ARIMA, LSTM, etc.) to model and predict the electrical parameter data.The ARIMA model is suitable for dealing with linear trends in time series data, while the LSTM (Long Short-Term Memory network) is a deep learning model that can capture non-linear trends in electrical parameter data and is especially suitable for predicting long-term trends. Based on historical electrical parameter data and trend prediction algorithms, the system can predict the future state changes of the device. For example, the system can predict the change trends of device current and voltage, discover potential faults in advance, and take preventive measures. Combining with image data (such as device monitoring videos or images), using image processing techniques (such as edge detection, image segmentation, etc.) to extract the characteristics of possible device failures. Through image analysis, external faults or abnormal conditions of the device can be visually identified, such as cable overheating, device damage, etc. Using machine learning algorithms (such as support vector machines, decision trees, neural networks, etc.) to learn and train the fault characteristics, so as to achieve intelligent diagnosis. The system can automatically diagnose the current device fault type based on the combination of historical fault patterns and image data, and give repair suggestions.

[0099] The effects of the above technical solution are as follows: By reasonably selecting storage devices (such as enterprise hard drives, solid-state drives, or cloud storage services) and configuring a RAID redundant array, not only can the reliability of data storage be improved, but the scalability of the system can also be enhanced, ensuring that as the data volume grows, the storage system can operate stably and implement an efficient data backup and fault tolerance mechanism; Using a database management system (DBMS) to store data in a normalized manner can achieve efficient data management. Through an indexing mechanism based on timestamps and device IDs, etc., the query and retrieval speed of data is significantly improved. Especially when processing a large amount of electrical parameter data and image frames, it can quickly respond to query requests and improve data access efficiency; Establishing a distributed storage and redundant backup mechanism ensures that data will not be lost even if a storage node fails. This redundancy mechanism greatly improves data security, ensures the high availability and reliability of the system, and is especially suitable for device monitoring and analysis systems that require long-term stable operation; Through programming tools such as Python to accurately parse and statistically analyze electrical parameter data, the operating status of the device can be monitored in real time. Based on the analysis results of the data, the system can promptly detect abnormal changes in parameters such as current and voltage, thereby predicting potential device failures or performance problems, reducing the workload of manual inspections, and improving device management efficiency; Using trend prediction algorithms (such as ARIMA, LSTM) to predict the device status can predict in advance the possible abnormalities or failures that the device may encounter in the future, providing an early warning for device maintenance. This prediction ability helps to carry out preventive maintenance, reduce the device failure rate, and reduce the risk of sudden downtime, thereby extending the service life of the device; Combining the analysis of image data and machine learning technology, the system can perform more intelligent fault diagnosis. When a device fails, it can extract fault features through image processing technology, compare them with known fault patterns, automatically identify the fault type, and give a repair plan. This intelligent fault diagnosis not only improves the accuracy of fault handling, but also reduces the error of manual judgment and improves the fault response speed; The overall solution constructs a complete automated monitoring and fault diagnosis system by combining various links of data storage, management, analysis, prediction, and diagnosis. The system can not only monitor the operating status of the device in real time, but also perform automatic alarm and prediction according to the analysis results, reducing manual intervention and improving the efficiency and automation level of overall device management.

[0100] In one embodiment of the present invention, the S35 includes:

[0101] Before trend prediction, preprocess the electrical parameter data and extract key features from the preprocessed data;

[0102] According to the characteristics of the data and the prediction requirements, select a trend prediction model. After selecting the model, use historical electrical parameter data for model training, and optimize the prediction performance by adjusting model parameters;

[0103] After the model training is completed, use the validation dataset to validate the model and evaluate its prediction accuracy and generalization ability; if the prediction result is not satisfactory, optimize the model.

[0104] Use the trained model to perform trend prediction on the electrical parameter data to obtain predicted values for a future period of time; at the same time, based on the anomaly detection mechanism, detect potential anomalies by comparing the differences between the predicted values and the actual values.

[0105] If an anomaly is detected, immediately trigger the fault warning mechanism to send warning messages to relevant personnel. At the same time, combine historical data and the expert database to provide fault handling suggestions for decision-makers.

[0106] The working principle of the above technical solution is as follows: Before performing trend prediction, it is first necessary to preprocess the collected electrical parameter data. This stage includes:

[0107] Data cleaning: Remove outliers from the data, such as extremely unreasonable current and voltage values, or untrue data introduced due to equipment failures and data acquisition errors; fill in missing values, and common methods include mean filling, interpolation, etc.

[0108] Data normalization or standardization: To improve the stability and accuracy of model training, it is usually necessary to perform standardization or normalization on data with different dimensions or ranges. For example, use Z-score standardization (i.e., subtract the mean from the data and divide by the standard deviation) or min-max normalization (scale the data to the range of 0 to 1).

[0109] Feature extraction: Extract key features from the preprocessed data. These features can be the inherent attributes of time series such as the periodicity, trend, and seasonality of the data. In addition, it is also necessary to analyze the correlation with other relevant electrical parameters (such as voltage, current, power, etc.); these features can be extracted through methods such as calculating the correlation coefficient and time series analysis.

[0110] Select a suitable trend prediction model according to the characteristics of the data and prediction requirements. Common choices include:

[0111] ARIMA model: For data with obvious periodicity and trend, the ARIMA model is a classic and effective method. The ARIMA model can capture the autocorrelation and differential trend in the data and is suitable for stable and linear time series.

[0112] LSTM neural network: For non-linear and complex time series data, deep learning models, especially LSTM (Long Short-Term Memory Network), can handle long-term dependence problems and are suitable for predicting complex and long-term electrical parameter data.

[0113] The training process includes:

[0114] Model training: Use historical electrical parameter data for model training. During the training process, adjust various parameters of the model (such as the p, d, q parameters in the ARIMA model, or the number of layers, number of units, learning rate, etc. in the LSTM network) to optimize the model's prediction ability;

[0115] Cross - validation and optimization: Use a validation dataset other than the training data to validate the model; Test its generalization ability by evaluating indicators such as the prediction accuracy, loss function, goodness of fit, etc. If the prediction accuracy is not ideal, model optimization is required, including adjusting hyperparameters, changing the network structure, etc.;

[0116] Use the trained and optimized model to predict the trend of electrical parameter data for a period of time in the future. The model will output predicted values, which reflect the future trend of electrical parameter changes of the device or system; During the prediction process, establish an anomaly detection mechanism to identify potential anomalies by comparing the difference between the actual value and the predicted value. If the deviation between the actual value and the predicted value exceeds a preset threshold, an anomaly is considered to have occurred.

[0117] The methods of anomaly detection include:

[0118] Statistical - based methods: Such as calculating the standard deviation of the prediction error and determining points with errors exceeding a certain multiple as anomalies.

[0119] Machine - learning - based methods: Anomaly detection models can be trained to identify potential abnormal data.

[0120] Once an abnormal situation is detected, the system will immediately trigger a fault warning mechanism. The warning information includes:

[0121] The type of anomaly (such as too high or too low voltage, abnormal current fluctuation, etc.);

[0122] Possible causes of the anomaly (for example, sudden increase in load, equipment failure, power supply fluctuation, etc.);

[0123] Suggested countermeasures (such as adjusting voltage, current load, checking equipment, enabling backup power, etc.).

[0124] The system combines historical data, equipment failure modes, and an expert database to provide detailed fault handling suggestions for decision-makers. These suggestions are based on similar events in historical data and combine expert experience to provide actionable steps for handling abnormal situations, thereby accelerating the fault repair process. As the data volume continues to increase and the system operation provides continuous feedback, the model should be retrained or fine-tuned regularly to ensure its adaptability to new data. By continuously optimizing the prediction model and adjusting the early warning mechanism, the system can continuously improve the prediction accuracy and fault response speed during actual operation.

[0125] The effects of the above technical solution are as follows: By comprehensively preprocessing the electrical parameter data (such as data cleaning, normalization, and feature extraction) and selecting an appropriate trend prediction model (such as ARIMA or LSTM), the characteristics such as periodicity, trend, and seasonality in the data can be fully utilized. This helps to improve the accuracy of trend prediction, reduce prediction errors, and enhance the predictability of the system; Through accurate trend prediction and anomaly detection mechanisms, the system can identify electrical parameter fluctuations or potential faults in advance and issue early warnings, thereby avoiding equipment failures or system downtimes. The timely fault early warning mechanism can ensure that relevant personnel can take countermeasures before problems occur, thus avoiding major losses; Through real-time monitoring and anomaly detection, the technical solution can quickly trigger an early warning when electrical parameters are abnormal and provide targeted fault handling suggestions based on the prediction model and historical data. This can significantly improve the speed and accuracy of fault response, reduce repair time and costs; The system continuously learns and optimizes from data during actual application. Over time, through continuous training, fine-tuning, and optimization, the model can adapt to new data characteristics and changes and maintain a high prediction accuracy. The continuous optimization process ensures that the system can adapt to the changing environment and improve long-term reliability; The system can not only issue an early warning when an anomaly is detected but also provide detailed fault handling suggestions based on historical data and the expert database. These suggestions help decision-makers react quickly, reduce dependence on manual intervention, and improve the efficiency and quality of fault handling; Since the system can predict and detect anomalies in real time, it reduces the dependence on manual monitoring. The intelligent anomaly detection and prediction mechanism effectively reduces the risk of manual misjudgment or neglecting potential faults and improves the operation and maintenance safety of the overall system; By predicting and detecting equipment failures in advance, the system can help maintenance personnel perform preventive maintenance before equipment failures occur. This not only extends the service life of the equipment but also greatly reduces the repair and replacement costs caused by sudden equipment failures; The technical solution is not only applicable to the electrical parameter monitoring in industries such as power and electricity but can also be extended to other fields according to actual needs, such as energy management and intelligent manufacturing, with strong adaptability and scalability.

[0126] In one embodiment of the present invention, S36 includes:

[0127] S361. Before analyzing in combination with the image data, preprocess the image data and extract the fault features from the preprocessed image;

[0128] S362. Match the extracted features with the known fault mode library to identify possible fault types. At the same time, use machine learning algorithms to automatically classify and identify the fault features;

[0129] S363. Combine the electrical parameter data and the image data to evaluate the severity of the fault; evaluate the severity of the fault and its impact on the equipment operation by calculating indicators such as the change rate and duration of the fault features; based on the evaluation results, establish a risk warning mechanism and send out warning signals in a timely manner;

[0130] S364. According to the fault diagnosis results and the risk warning information, formulate targeted preventive maintenance strategies, record and organize the cases of each fault diagnosis and maintenance, and build a fault case library.

[0131] The working principle of the above technical solution is as follows: First, denoise the original image to remove the noise in the image. This can be achieved through common image denoising algorithms such as median filtering and mean filtering. The purpose of denoising is to improve the quality of the image and make subsequent processing more accurate; enhance the details and clarity of the image through contrast enhancement techniques (such as histogram equalization, local contrast enhancement, etc.). At the same time, correct the color of the image to ensure that the colors in the image are consistent with the actual situation and reduce color deviations caused by lighting or shooting angles; for blurred images, image sharpening techniques (such as Laplacian filtering, Sobel operator, etc.) can be used to improve the clarity for better observation of the details in the image. Sharpening can highlight the edges of the image, enhance the structural information, and make potential fault features more prominent; for dynamic images, perform frame synchronization processing to ensure the alignment and time synchronization between consecutive frames. This step is very important for accurately identifying dynamic faults and avoiding misidentification due to motion blur or inconsistent frame rates; extract fault features from the preprocessed image by applying image processing algorithms such as edge detection (such as Canny edge detection), image segmentation, morphological processing (such as dilation and erosion), etc. The extracted features include abnormal shapes, color changes, texture differences, etc., and these features are important bases for fault diagnosis; match the extracted features with a known fault pattern library. The fault pattern library contains the typical features of the historical faults of the equipment, including the visual manifestations of different types of faults. Through matching, the possible fault types in the image can be initially identified; use machine learning algorithms such as convolutional neural networks (CNNs) to automatically classify and identify the fault features. CNNs can automatically learn the features of different types of faults from a large amount of image data and then accurately classify new images. This automated classification method can improve the efficiency and accuracy of fault identification and reduce manual intervention; combine electrical parameter data (such as current, voltage, temperature, etc.) with image data for a more comprehensive analysis of the faults. Electrical parameter data provides the real-time operating state of the equipment, while image data provides abnormal information about the appearance of the equipment; evaluate the severity of the fault by analyzing the change rate (such as the speed of fault expansion, the rate of morphological change, etc.) and duration (the time span of fault occurrence) of the extracted fault features. These indicators help to determine whether the fault is in the development stage and whether it will have a long-term impact on the equipment; based on the severity assessment results of the fault, the system can establish a risk warning mechanism. When the system identifies that the fault reaches a certain severity threshold, it issues a warning signal in a timely manner to notify relevant personnel to take emergency measures. This helps to reduce equipment downtime and avoid greater losses; formulate targeted preventive maintenance strategies according to the fault diagnosis results and risk warning information. Preventive maintenance includes operations such as regular inspections, replacement of key components, and adjustment of equipment parameters.Through these measures, the occurrence of faults can be minimized, and the service life of the equipment can be extended; during each fault diagnosis and maintenance, relevant cases will be recorded and sorted to build a fault case library. The fault case library includes detailed information such as fault phenomena, diagnosis processes, maintenance measures, and results. By accumulating a large number of fault handling cases, the fault diagnosis and maintenance processes can be continuously optimized, providing valuable experience and reference for future fault handling.

[0132] The effects of the above technical solutions are as follows: Through the preprocessing of image data (such as denoising, enhancing contrast, image sharpening, etc.) and feature extraction (such as edge detection, image segmentation, etc.), the manifestations of equipment faults can be identified more clearly, thereby improving the accuracy of fault detection. Combining machine learning algorithms (such as CNN) to automatically classify and identify fault features further improves the efficiency of fault identification, reduces the risk of manual intervention and misidentification; combining image data with electrical parameter data for fault severity assessment can comprehensively consider the appearance anomalies and real-time operation data of the equipment. This multi-dimensional analysis method can not only accurately assess the severity of faults but also comprehensively understand the impact of faults on equipment operation, helping to make more accurate maintenance decisions; based on the fault severity assessment, the established risk warning mechanism can issue fault warnings in a timely manner, reduce the downtime of the equipment, and avoid sudden faults. The warning signal provides valuable advance preparation time for equipment maintenance personnel, helping to take measures in advance and reduce potential losses; according to the fault diagnosis results and warning information, the formulated preventive maintenance strategy can targetedly solve potential problems and improve the stability and long-term reliability of the equipment by means of regular inspections, replacing key components, adjusting equipment parameters, etc. This strategy helps to reduce the frequency of fault occurrence, extend the service life of the equipment, and reduce maintenance costs; through the establishment of a fault case library, the experience of each fault handling has been sorted and recorded, providing data support and reference for future fault diagnosis and maintenance. This knowledge accumulation continuously optimizes the fault diagnosis process and improves the intelligent level of the system. As the case library expands, the system can better handle new types of faults and improve the overall operation and maintenance efficiency; through automated image processing and machine learning classification, the need for manual intervention is reduced, thereby avoiding misjudgment or omission caused by human factors. This not only improves the accuracy of fault diagnosis but also reduces labor costs and improves the overall operation and maintenance efficiency of the system; by detecting faults early and taking preventive measures in a timely manner, the fault downtime of the equipment can be effectively reduced, and the impact of production line stagnation or system faults on business can be avoided.

[0133] In an embodiment of the present invention, the S362 includes:

[0134] On the basis of the preliminary extraction of fault features, further refine the feature extraction process; utilize image processing techniques, such as the Feature Pyramid Network in deep learning or super-resolution reconstruction technology, to enhance and extract the minute details in the images, and obtain a more refined and comprehensive fault feature set;

[0135] Preprocess the extracted fault features, and at the same time, standardize the features;

[0136] Based on historical fault data and expert experience, construct a known fault mode library; at the same time, assign a unique identifier to each fault type;

[0137] Through the dynamic update mechanism of the fault mode library, continuously update and improve the fault mode library according to newly emerging fault cases and expert feedback; at the same time, use machine learning algorithms to automatically learn and summarize the fault modes;

[0138] Use a similarity measurement algorithm to match the extracted fault features with the features in the known fault mode library; initially identify possible fault types by calculating the similarity scores between the features;

[0139] According to the feature matching scores, screen and rank the initially identified fault types; for the fault types with higher scores, consider them as key objects; for the fault types with lower scores, conduct further analysis and verification;

[0140] Use machine learning algorithms to train the features in the known fault mode library and construct an intelligent classification and recognition model;

[0141] Verify and optimize the output results of the intelligent classification and recognition model; use cross-validation to evaluate the performance of the model; adjust and optimize the model according to the evaluation results.

[0142] The working principle of the above technical solution is as follows: First, the system extracts the preliminary features of the fault through image processing technology, including abnormal shapes, color changes, texture differences, etc. This process can capture the visual abnormalities on the surface of the device or its working state. Next, deep learning methods such as Feature Pyramid Network (FPN) are used to further process the image, enhancing the tiny details in the image to obtain more refined and comprehensive fault features. This can effectively improve the ability to identify complex and tiny faults. In addition, super-resolution reconstruction technology can be used to improve the resolution of the image, helping to more clearly identify the tiny details in the image. The extracted features are subjected to a series of preprocessing operations, including feature scaling, normalization, etc., to make the feature data numerically consistent and stable. This helps subsequent algorithm processing and improves classification accuracy. Further standardize the features so that each feature dimension has the same scale and distribution to avoid certain features dominating the learning process of the model due to scale differences. By combining historical fault data and the experience of domain experts, a fault mode library is constructed. This library contains information such as typical features, fault manifestations, causes, and solutions for various fault types. Each fault type has a unique identifier in the library for subsequent matching and classification. As new fault cases emerge and expert feedback is updated, the fault mode library will be dynamically updated. This ensures that the system can continuously adapt to new fault modes and improve accuracy and adaptability over time. Similarity measurement algorithms (such as cosine similarity, Euclidean distance, etc.) are used to match the extracted fault features with the features in the fault mode library. Calculate the similarity scores between the features to initially identify the fault type. The fault type with a higher score will be used as the main candidate type, and the types with lower scores need further analysis. According to the scores of feature matching, the initially identified fault types are screened and sorted. The fault types with higher scores will be given priority and further verified and processed. The types with lower scores need to be further confirmed as the fault source through more analysis (such as on-site inspection or additional data support). Machine learning algorithms (such as Convolutional Neural Network CNN, Support Vector Machine SVM, etc.) are used to train the features in the fault mode library to establish an intelligent classification and recognition model. This model can automatically classify and recognize newly extracted fault features and output the possible fault types and their confidence levels. The output results of the model need to be verified through methods such as cross-validation to evaluate its accuracy and reliability. According to the evaluation results, optimize and adjust the parameters of the model to improve its performance and recognition ability. During the process of fault mode recognition, the output of the model is verified and optimized to ensure that the identified fault types are accurate. Evaluate the performance of the model using evaluation metrics such as cross-validation, accuracy, recall rate, etc., and adjust and optimize the model according to the evaluation results to ensure that the model can adapt to various new fault scenarios.

[0143] The effects of the above technical solutions are as follows: By using the Feature Pyramid Network (FPN) and super-resolution reconstruction technology in deep learning, the tiny details in the images can be effectively enhanced, so as to extract more accurate and comprehensive fault features. This refined feature extraction can better capture the subtle manifestations of complex faults, thereby improving the accuracy of fault detection; the steps of feature scaling, normalization, and standardization make the fault features consistent in terms of value and structure, thus improving the stability and classification accuracy of the subsequent model. The standardization operation of feature preprocessing can eliminate the scale differences between different features, enabling the model to better classify and analyze; by combining historical fault data and expert experience, a fault mode library containing various typical fault features, manifestations, and solutions is established. This not only provides multi-dimensional information support but also enables the system to continuously learn new fault modes through a dynamic update mechanism, enhancing its ability to handle new types of faults; by matching the fault features through a similarity measurement algorithm, the fault type can be initially identified, and key analysis can be carried out based on the score ranking. This automated identification and screening process greatly reduces the workload of manual analysis while improving the efficiency and accuracy of fault diagnosis; by training with machine learning algorithms such as Convolutional Neural Network (CNN) and Support Vector Machine (SVM), an intelligent classification and recognition model can be constructed, enabling the system to automatically process and classify fault features. This intelligent recognition model can not only provide accurate fault type recognition results but also evaluate the confidence of each prediction result, further reducing the risks of misidentification and missed identification; through cross-validation and performance evaluation, the model can be continuously adjusted and optimized, thereby improving the recognition accuracy. This optimization mechanism ensures that the model can adapt to different application scenarios and continuously changing fault types, maintaining a high recognition rate and reliability.

[0144] In one embodiment of the present invention, step S4 includes:

[0145] S41. Quickly locate to a specific time period through the time code to narrow the scope of fault troubleshooting, and retrieve relevant data packets in the database using the timestamp query function;

[0146] S42. Compare the electrical parameter data and image materials within this time period, analyze the changes in the device state before and after the fault occurs, and use a data visualization tool to display the change trend of the electrical parameter data;

[0147] S43. Conduct a comprehensive analysis of the fault by combining the electrical parameter data and image materials, and use multivariate analysis technology to identify the fault features;

[0148] S44. Determine the specific cause and scope of influence of the fault, conduct a fault cause analysis using fault tree analysis, and based on the analysis results, propose repair suggestions or preventive measures;

[0149] S45. Generate a detailed fault report based on the fault analysis results, format the fault report using a report generation tool, add charts, pictures, and explanations, and store and share the fault report with relevant personnel.

[0150] The working principle of the above technical solution is as follows: Through the time code information in the system, quickly locate the specific time period when the fault occurs. The time code (such as a time stamp) is an accurate time identifier recorded by the system when the fault occurs. Through this time code, the troubleshooting scope can be accurately narrowed to avoid blind searching; once the specific time period is located, relevant data packets (such as electrical parameter data, device status data, image materials, etc.) within this time period can be retrieved from the database through the time stamp to ensure the comprehensiveness and accuracy of the fault analysis data; Next, compare the electrical parameter data and image materials within this time period. Electrical parameter data (such as voltage, current, power, frequency, etc.) can reflect the operating state of the device, while image materials (such as pictures or videos taken by the device's camera) can provide an intuitive change in the device's appearance or working environment; Use data visualization tools, such as line charts, scatter plots, etc., to display the change trend of the electrical parameter data. Through intuitive charts, abnormal fluctuations before and after the fault occurrence can be quickly identified, providing data support for subsequent fault analysis; Combine the electrical parameter data and image materials for comprehensive analysis. By integrating the two data sources, more comprehensive and accurate fault information can be obtained. For example, sudden changes in electrical parameters may be related to factors such as changes in the device's appearance and environmental conditions; Through principal component analysis, reduce the dimensionality of high-dimensional data, extract the most important fault features, remove noise, and find the most representative fault patterns in the data; According to the similarity of the data for clustering, data of similar fault types can be grouped into one category, so as to identify whether there are potential hidden dangers or abnormal trends in the device; Use fault tree analysis technology to deeply analyze the cause of the fault. Through FTA, a hierarchical fault model can be established to clarify the root cause of the fault occurrence, and layer by layer trace back the possible influencing factors and potential problems; If the FTA analysis shows that the failure of some key components leads to the fault, it is necessary to recommend maintenance personnel to replace these components; If the operating parameters of the device deviate, the working state or operating conditions of the device need to be adjusted to ensure its normal operation; Specific repair plans should be proposed for each step in the fault analysis process in combination with the specific device, working environment, and fault type; Through the results of the fault analysis, the system will automatically generate a detailed fault report. The report content includes:

[0151] Fault type, location, cause, and scope of influence;

[0152] Repair suggestions or preventive measures (such as replacing components, adjusting parameters, etc.);

[0153] Auxiliary information such as data visualization charts, image materials, etc.;

[0154] Use reporting tools (such as Word, Excel, etc.) to format the analysis results for easy viewing and sharing. This process may include inserting charts, data analysis results, repair suggestions, etc., to make the report more intuitive and understandable; the generated report will be stored in the system and can be shared with relevant personnel, such as equipment maintenance personnel, technicians or management, via email, sharing platforms or other tools. This link ensures that fault information can be quickly transmitted and appropriate actions can be taken.

[0155] The effects of the above technical solutions are as follows: By quickly positioning the time period when the fault occurred through the time code, the troubleshooting scope can be rapidly narrowed, avoiding the cumbersome and ineffective searches in traditional troubleshooting methods, thus saving a large amount of time and resources; Using the timestamp query function can accurately obtain relevant data packets, ensuring the pertinence and accuracy of data analysis and reducing the error of manual judgment; Comparing the electrical parameter data with the image data can comprehensively understand the state changes of the equipment before and after the fault occurs. Data visualization tools (such as line charts, scatter plots, etc.) make the trend of electrical parameter changes clear at a glance, facilitating the identification of abnormal fluctuations during equipment operation and ensuring that fault characteristics are clearly visible; By combining electrical parameter data with image data and using multivariate analysis techniques (such as principal component analysis, clustering analysis, etc.), the key characteristics of the fault can be extracted from complex data, avoiding misjudgment that may be caused by a single data source; Using fault tree analysis (FTA) can help accurately identify the root cause and scope of influence of the fault and gradually trace potential fault sources. This method is systematic and in-depth, ensuring the comprehensiveness and accuracy of fault analysis; Through the analysis of FTA, it can be ensured that the handling of equipment faults is more precise, providing more targeted guidance for repair and maintenance work and reducing ineffective or incorrect repair solutions; Combining the fault analysis results to put forward specific repair suggestions or preventive measures, such as replacing damaged components, adjusting equipment parameters, etc., can greatly improve the accuracy and efficiency of maintenance work; Maintenance personnel can reduce the diagnosis time and unnecessary component replacement according to the detailed report and clear fault cause, reducing the maintenance cost; By generating a detailed and formatted fault report, the fault type, cause, scope of influence and repair suggestions can be clearly displayed, making the report easier to understand and convey; The reporting tools for fault reports (such as Word, Excel, etc.) make the report content neat, intuitive and can be conveniently shared electronically with relevant personnel (such as equipment maintenance personnel, technicians, management, etc.), accelerating the communication and decision-making process for fault handling; Systematic fault analysis and report generation contribute to the establishment of an equipment fault database, providing the accumulation and analysis support of historical data and providing valuable data references for subsequent equipment maintenance and optimization; During the operation of the equipment, it can more effectively track and manage the equipment status, contributing to preventive maintenance and fault prediction, extending the service life of the equipment and reducing the impact of sudden faults on production.

[0156] An embodiment of the present invention, said S43, includes:

[0157] Integrate the electrical parameter data and the image data to ensure their synchronization in time, including timestamp alignment of the electrical parameter data with the image data;

[0158] Clean the integrated data and perform standardization processing on the data;

[0159] Apply principal component analysis technology to reduce the dimensionality of the electrical parameter data and extract the main components affecting the change of the device state; Based on clustering analysis technology, divide the electrical parameter data and the image data into different categories or groups;

[0160] Analyze the correlation relationship between the electrical parameter data and the image data through association rule mining technology; Based on the results of multivariate analysis, select the features that make significant contributions to fault identification;

[0161] Use feature extraction technology to extract fault features from the electrical parameter data and the image data, and fuse the extracted electrical parameter features and image features to form a comprehensive fault feature vector;

[0162] Use machine learning or deep learning algorithms to perform pattern recognition on the fused fault feature vector, and based on the results of pattern recognition, perform a detailed diagnosis of the fault.

[0163] The working principle of the above technical solution is as follows: The electrical parameter data and image materials need to be strictly aligned according to the time stamp. The electrical parameter data is usually collected regularly, such as current, voltage, power, etc., and the image materials are collected through video monitoring or infrared thermal imaging equipment. Through time stamp alignment, each frame of the image can be corresponding to the electrical parameter data, thus providing an accurate time correlation for subsequent analysis; Align the electrical parameter data and image materials (such as video frames, thermal imaging images, etc.) in the time dimension to ensure that the electrical parameter data at each time point can be matched with the relevant image materials, so as to obtain more comprehensive equipment status information; After data integration, the data must be cleaned to remove outliers, missing values or invalid data to ensure the accuracy of the analysis results. For example, for the extreme values or unreasonable values in the electrical parameter data, interpolation methods or other statistical methods can be used for repair; The electrical parameter data needs to be standardized, usually processed by dimensionless (such as z-score standardization) to eliminate the influence of the dimension. The image materials need to be normalized to compress the pixel values into a fixed range (such as [at 0,1]) to eliminate the influence of brightness, contrast, etc. of different images, which is convenient for subsequent analysis; The electrical parameter data often contains multiple variables, and there may be a high degree of correlation between these variables. Through principal component analysis (PCA), the dimensionality of the electrical parameter data can be reduced, and the main components that have the greatest impact on the change of the equipment status can be extracted. The goal of PCA is to reduce the dimension of the data while retaining the most important information and reducing the complexity of the analysis; Cluster analysis is used to divide the electrical parameter data and image materials into different groups or categories according to similarity. Through cluster analysis, similar patterns or fault types in the same operating state can be identified, which helps subsequent fault diagnosis and prediction; Apply association rule mining techniques (such as Apriori algorithm, FP-growth, etc.) to mine the potential association relationships between the electrical parameter data and image materials. For example, it may be found that the changes of certain electrical parameters have a strong correlation with specific image features (such as equipment heating, surface cracks, etc.). Through these rules, the relationship between the changes of electrical parameters and equipment faults can be revealed, helping to understand the potential causes of equipment anomalies; For the electrical parameter data, various statistical quantities can be calculated, such as mean, variance, skewness, kurtosis, etc. These statistical quantities can reflect key information such as the stability and fluctuation of equipment operation; The feature extraction of image materials usually involves image processing techniques, such as edge detection, corner detection, texture analysis, etc. Through these methods, important features in the image can be extracted, such as the surface heat distribution of the equipment, appearance defects, wear conditions, etc.; Integrate the electrical parameter features and image features to form a comprehensive fault feature vector. In this way, the electrical data can be combined with the data of the equipment appearance change, so as to more comprehensively describe the status and potential faults of the equipment; Using the fused feature vector, the data can be trained through machine learning or deep learning models (such as support vector machines, decision trees, neural networks, etc.) for pattern recognition.These algorithms can learn from known failure modes and predict new types of failures; based on the results of pattern recognition, the algorithms can conduct detailed diagnosis of equipment failures. Specifically, the algorithms can analyze the changes in electrical parameters and abnormal features in the image data to determine the specific causes of the failures (such as equipment aging, electrical faults, external environmental impacts, etc.), and evaluate the scope of the impact of the failures (such as the degree of equipment damage, production downtime, etc.); through fault diagnosis, the system can estimate the degree of impact of the failures on the equipment, including the degree of damage, repair difficulty, production losses, etc. This information helps to formulate maintenance and emergency response strategies; based on the fault diagnosis results, the system can provide specific repair suggestions or preventive measures. For example, if the failure is caused by overload, it may be recommended to reduce the load or replace the faulty components; if the failure is caused by too high environmental temperature, it may be recommended to strengthen the cooling system.

[0164] The effects of the above technical solution are as follows: By integrating electrical parameter data and image materials and adopting technologies such as principal component analysis and cluster analysis, the changes in the operating state of the device can be accurately identified. By extracting the fused electrical parameters and image features, the working conditions of the device can be comprehensively reflected, thereby improving the accuracy of fault identification. With the help of machine learning and deep learning algorithms, the system can automatically learn and identify fault patterns, significantly improving the efficiency of fault diagnosis, reducing manual intervention and human errors; by combining the analysis of electrical parameters and image materials, the system can predict potential faults in advance, especially when the device is abnormal. Through the real-time monitoring and analysis of the characteristic data, the system can issue early warnings before the occurrence of faults, helping maintenance personnel to intervene and repair in time, and avoiding equipment downtime or serious damage. This preventive maintenance method can effectively reduce the equipment downtime and unnecessary maintenance costs; through the integrated analysis of electrical parameter data and image materials, the health status of the device can be comprehensively evaluated from multiple dimensions. The electrical parameter data provides information on the operating state of the device, while the image materials can reveal changes in the appearance, surface condition, and temperature of the device. The combination of the two makes fault diagnosis not limited to the analysis at the electrical level, but to identify potential problems from a more comprehensive perspective and provide more accurate fault causes; by using cluster analysis and pattern recognition technologies, the system can automatically classify different device states and fault types, effectively distinguishing normal states from various fault patterns. This enables maintenance personnel to locate problems more quickly and take targeted measures according to different fault types, improving the speed and quality of fault response; through accurate fault prediction and diagnosis, equipment maintenance can be transformed from "repair after the event" to "preventive maintenance". This transformation not only reduces the risk of equipment damage but also optimizes the allocation of maintenance resources, avoiding unnecessary maintenance work and component replacement, thereby significantly reducing the overall maintenance cost; by continuously monitoring and analyzing the device state, combining electrical parameter data and image materials, the system can timely detect minor abnormalities of the device and conduct early intervention. This timely detection and repair can prevent the expansion of faults and further damage to the device, thereby extending the service life of the device and improving the overall reliability of the device; this solution not only provides means for fault diagnosis but also provides data support for device operation, helping managers better understand the operating state and health status of the device. Through these data, the maintenance strategy, usage plan, and resource allocation of the device can be better optimized and adjusted, thereby improving the overall operating efficiency of the device; this technical solution provides a data-driven decision support system for device management and maintenance. By continuously analyzing electrical parameters and image materials, the system can provide a basis and guidance for device management, helping enterprises make more well-grounded repair and investment decisions, and avoiding blind repairs or ignoring potential problems.

[0165] In one embodiment of the present invention, the S5 includes:

[0166] S51. Select appropriate visualization forms according to different fault diagnosis requirements, perform visualization processing on the extracted electrical parameter data and image materials, and use a data visualization library to generate intuitive charts, curves, animations, etc.;

[0167] S52. Optimize and adjust the visualization results, add annotations, labels or legends as needed; and display the visualization results to relevant personnel.

[0168] The working principle of the above technical solution is as follows: According to specific fault diagnosis requirements, first determine the visualization goal of the data. For example, if it is necessary to show the change trend of electrical parameter data, select an appropriate chart type, such as a line chart, which can clearly show the change of data over time; if it is necessary to simulate the fault occurrence process, animations may be needed to vividly show the state changes of the device before and after the fault. Obtain electrical parameter data and image materials from the device. The electrical parameter data includes the current, voltage, power, etc. of the device, while the image materials contain monitoring information such as the appearance, temperature, vibration, etc. of the device. These data need to be processed through certain cleaning and standardization for subsequent visualization display; use mature data visualization libraries, such as Matplotlib, Plotly, Seaborn, etc., and select appropriate tools according to specific requirements. For example:

[0169] Matplotlib: Suitable for static graphics, often used for line charts, bar charts, etc. to display electrical parameter data;

[0170] Plotly: Suitable for dynamic interactive graphics, capable of creating three-dimensional graphics that can be scaled and rotated, especially suitable for showing complex dynamic processes;

[0171] Seaborn: Used to display more complex statistical graphics, such as heat maps, regression plots, etc.

[0172] Generate charts or animations by selecting appropriate visualization forms. For line charts, Matplotlib can be directly used to plot the time series of electrical parameter data; for dynamic fault simulation, animations can be used to simulate and combine Plotly or other tools to show the behavior changes of the device under different fault scenarios; after generating the visualization results, the details of the charts need to be adjusted according to requirements, for example:

[0173] Adjust colors: According to the complexity and presentation method of the chart content, select a reasonable color scheme so that different data categories can be clearly distinguished.

[0174] Line thickness: For line charts, bar charts, etc., adjusting the line thickness can highlight important data trends and ensure that key parts are fully displayed.

[0175] Font size: Adjust the font size according to the display resolution and the needs of the audience to ensure the readability of the chart.

[0176] Add notes and labels: Add detailed notes and labels to the chart to explain the specific meaning of the data or indicate a certain special time point (such as the moment of failure).

[0177] Add a legend: For data containing multiple data lines or multiple categories, a legend can be added to help users quickly identify different data categories.

[0178] Present the finally optimized chart, curve or animation to the relevant personnel. The presentation methods can be:

[0179] Report or presentation: Insert the visualization results into a report or PPT to help technicians, equipment maintenance personnel or management understand the status of the equipment and potential failure risks.

[0180] Real-time monitoring system: Embed the chart into a real-time monitoring platform to provide real-time fault diagnosis and trend monitoring for staff to check at any time.

[0181] The effects of the above technical solutions are as follows: By selecting appropriate visualization forms according to fault diagnosis requirements, technicians can more intuitively analyze and understand the operating status of equipment. Using line charts to display the change trends of electrical parameter data helps to quickly identify abnormal fluctuations or precursors of faults in equipment; By animating the fault occurrence process, the performance of the equipment under different conditions can be clearly presented, thus improving the fault diagnosis efficiency; Through data visualization processing, converting electrical parameter data and image materials into forms such as charts, curves, or animations enables relevant personnel to more easily understand complex equipment operation data. This intuitive presentation helps technicians, equipment maintenance personnel, and management make accurate judgments, avoid missing important information, and thus improve the accuracy of decision-making; The visualization results can clearly convey the status and fault conditions of the equipment, enabling personnel at different levels (such as technicians, management, etc.) to communicate more effectively and avoid misunderstandings or decision-making mistakes caused by information asymmetry. In addition, adding annotations, labels, and legends helps to further explain the specific meaning of the data, thus promoting teamwork; The optimized visualization charts or animations can be embedded in the real-time monitoring system to provide technicians with real-time updates on the equipment status, helping them identify potential faults in advance and take corresponding measures. This real-time monitoring and fault prediction ability can significantly improve the equipment maintenance efficiency, reduce downtime, and extend the service life of the equipment; By optimizing and adjusting the visualization results (such as adjusting colors, line thicknesses, font sizes, etc.), a clearer and more understandable graphical interface can be provided, avoiding information being too complex or difficult to understand and enhancing the user experience. Whether it is a report, presentation, or real-time monitoring system, it can effectively help users obtain the key information they need.

[0182] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these changes and modifications.

Claims

1. A fault diagnosis and recording method based on images and data, characterized in that: The method comprises: S1. Real-time collection of electrical parameter data and operating status image data; S2, combining the time code with the collected electrical parameter data and image data; S3. storing the encapsulated data packets through a data storage device; and parsing and analyzing the stored data packets; S4. Accurately diagnose and locate faults; S5. Display through visual form.

2. The fault diagnosis and recording method based on images and data according to claim 1, characterized in that: Said S1 comprises: S11, select sensor; S12, configure sensor data collection frequency; S13. Deploy high-definition cameras at key locations on the equipment; S14, performing preliminary processing on the collected data; S15. Transmit the processed data to the data processing center in real time.

3. The fault diagnosis and recording method based on images and data according to claim 2, characterized in that: Said S15 comprises: Compress the processed data using a data compression algorithm; Dynamically select the optimal transmission channel and perform parallel data transmission; Dynamically adjust the sending order and size of data packets; Add an accurate timestamp to each data packet to keep the clock synchronized; Real-time monitoring of relevant performance indicators of each transmission channel; Through the feedback mechanism, targeted adjustments and optimizations are made.

4. The fault diagnosis and recording method based on images and data according to claim 1, characterized in that: Said S2 comprises: S21, synchronizing the time code with the collected electrical parameter data and image data; S22, verifying the synchronized data; S23, encrypting the data packet; S24. Store the checksum together with the data packet, and check the data packet again before storing.

5. The fault diagnosis and recording method based on images and data according to claim 4, characterized in that: Said S21 comprises: Generate continuous and stable time code based on the configured high-precision clock source; Distribute the generated time code to various data collection points and data processing centers through dedicated channels or networks; Align the time code with the timestamp of the data point; Preprocess the raw data; Real-time monitoring of time synchronization accuracy and stability; Corrective actions are taken automatically.

6. The fault diagnosis and recording method based on images and data according to claim 1, characterized in that: Said S3 comprises: S31. Select storage devices based on data volume and storage requirements, and perform performance testing on the storage devices. S32, storing the encapsulated data packets in a storage device in chronological order through a database management system for data storage and management; S33. Based on the data index mechanism, the stored data is backed up and redundancy processed through distributed storage; S34. Parse the data packet using Python to extract the timestamp, electrical parameter values, and image frame data, perform statistical analysis on the electrical parameters, and detect abnormal changes in the device's operating status from the analysis results. S35. Perform trend prediction on the electrical parameter data based on the trend prediction algorithm, and predict possible future state changes of the equipment; S36. Combine image data to conduct intuitive fault analysis and diagnosis, use image processing technology to extract fault features, and compare them with known fault patterns, while using machine learning algorithms for intelligent fault diagnosis.

7. The fault diagnosis and recording method based on images and data according to claim 6, characterized in that: The S35 includes: Before trend prediction, the electrical parameter data is preprocessed and key features are extracted from the preprocessed data; Select a trend prediction model based on the data characteristics and prediction requirements. After selecting the model, train the model using historical electrical parameter data and optimize the prediction performance by adjusting the model parameters. After the model training is completed, the model is verified using the validation dataset; Use the trained model to predict the trend of electrical parameter data and obtain the predicted value for a period of time in the future. At the same time, based on the anomaly detection mechanism, potential anomalies are detected by comparing the difference between the predicted value and the actual value. If an abnormal situation is detected, the fault warning mechanism will be triggered and warning information will be sent to relevant personnel. At the same time, combined with historical data and expert database, fault handling suggestions will be provided to decision makers.

8. The fault diagnosis and recording method based on images and data according to claim 6, characterized in that: The S36 includes: Before analyzing the image data, the image data is preprocessed and fault features are extracted from the preprocessed image based on the image processing algorithm; The extracted features are matched with a known fault pattern library to identify possible fault types, and machine learning algorithms are used to automatically classify and identify fault features. Combine electrical parameter data and image data to assess the severity of the fault; calculate indicators of fault characteristics to assess the severity of the fault and its impact on equipment operation; establish a risk warning mechanism based on the assessment results; Based on the fault diagnosis results and risk warning information, formulate targeted preventive maintenance strategies, record and organize each fault diagnosis and maintenance case, and build a fault case library.

9. The fault diagnosis and recording method based on images and data according to claim 1, characterized in that: Said S4 comprises: S41, using a timestamp query function to retrieve relevant data packets in a database; S42. Analyze the changes in equipment status before and after the fault occurs; S43. Combine electrical parameter data and image data to conduct comprehensive fault analysis; S44. Based on the analysis results, propose repair suggestions or preventive measures; S45. Store the fault report and share it with relevant personnel.

10. The fault diagnosis and recording method based on images and data according to claim 1, characterized in that: Said S5 comprises: S51, performing visualization processing on the extracted electrical parameter data and image data; S52. Optimize and adjust the visualization results, and present the visualization results to relevant personnel.

Images