124 results about "Offline analysis" patented technology

The invention discloses a dynamic monitoring and analyzing method for a thermal management component molten pool based on image recognition, and relates to the technical field of laser powder bed melting additive manufacturing. The method comprises the following steps: continuously capturing an original thermal management component molten pool image and temperature field distribution data during laser powder bed melting additive manufacturing; and optimizing the original thermal management component molten pool image to generate an optimized thermal management component molten pool image. Molten pool images and temperature field data are captured in real time through a high-frame-rate synchronous camera and an infrared sensor, defect types are recognized online in combination with a lightweight classification model, classification results are fed back to a control system to dynamically adjust laser power or scanning speed, interlayer defect accumulation can be inhibited in time, and the defect quality is improved. The defect diffusion problem caused by response lag of a traditional off-line analysis method is avoided, and the forming precision and the process stability of the heat management component are remarkably improved.

The invention discloses a phishing mail detection method and system based on a large model, and relates to the technical field of network security, the system comprises a real-time detection system and an offline analysis system, and a mail content analysis module receives real-time mail data to form structured data; the detection rule center matches the structured data with a detection rule, directly intercepts a mail hitting a high-confidence blacklist rule, and transmits the structured data of a mail hitting a low-confidence blacklist rule into a phishing mail detection agent; the phishing mail detection agent sequentially performs mail header suspicious feature analysis, mail body semantic analysis, mail body structure analysis and attachment content analysis on the structured data to obtain a mail intention, a link and an attachment file; and performing corresponding analysis by combining a detection tool, and judging whether the mail is a phishing mail or not according to an analysis result. A large-model-driven phishing mail detection intelligent agent is utilized, and a detection path is dynamically planned to cope with endless phishing mail attack means.

The application discloses a kind of electric power infrastructure line operation behavior intelligent identification method, belong to pattern recognition and data processing technical field, it includes constructing heterogeneous space-time scene atlas to carry out multi-modal structured perception to operation site;Continuous behavior flow is unsupervised and is divided into behavior base element and is encoded as observation sequence;Offline analysis safety operation regulation and is compiled as paradigm behavior genome;Observation sequence and paradigm genome are locally sequenced to be aligned to quantify behavior deviation;The jump critical point of risk state is identified based on deviation disturbance, and structured early warning event is generated.The application constructs heterogeneous space-time scene atlas to fuse multi-modal perception data, abstracts actual operation behavior as observation behavior base element sequence, and carries out sequence alignment with paradigm behavior genome compiled from safety regulation, can quantify behavior deviation and identify the jump critical point of risk state, realize accurate, forward-looking intelligent identification and early warning to operation behavior.

The invention discloses a natural resource big data off-line comprehensive analysis, annotation and proof system and method, and belongs to the technical field of natural resource informatization and geographic information processing. The system comprises the following components: a spatial data fusion module, which is used for realizing unified management and efficient processing of heterogeneous spatial data; the spatial index optimization module is used for realizing quick retrieval and high-performance query optimization of spatial data; the intelligent annotation and arrangement module is used for realizing low-code rapid configuration and business process automatic arrangement of different natural resource investigation scenes; the off-line analysis module is used for realizing buffer area analysis, space superposition and terrain analysis in an off-line environment; the real-time visualization module is used for realizing multi-level map display, track recording, vector element editing and multimedia evidence collection; and the data quality management module is used for realizing automatic quality inspection of geometry, topology and attributes and statistical analysis, version control and export management of annotation results.

The invention discloses a method for automatically inspecting equipment in intelligent manufacturing, which comprises the following steps of: 1, deploying edge computing nodes, combining with 5G network slices, cleaning and compressing inspection data in real time, dividing priorities, performing low-delay transmission on key data through a special bandwidth, and locally caching or filtering non-key data; 2, developing a controller based on the software defined network, monitoring a link state, introducing a reinforcement learning algorithm, and dynamically optimizing a transmission path; 3, deploying a TinyML framework at an equipment end, integrating a model compression technology to operate a lightweight AI model, and combining sensor data to perform real-time early warning and support offline analysis; 4, constructing an incremental learning framework, dynamically updating model parameters through knowledge distillation, and only adjusting part of parameters to adapt to state change; and step 5, an analysis result is fed back to the system by using a message queue, and a patrol strategy is dynamically adjusted in combination with a rule engine to form a closed-loop optimization mechanism of data acquisition-transmission-analysis-decision.

The invention discloses an edge reasoning optimization method and system based on segmented knowledge distillation. Firstly, a segmented knowledge distillation framework is constructed, a teacher-student model is divided into corresponding sub-modules with balanced parameters, parallel distillation training is adopted, middle feature dimension reduction and space alignment are achieved in combination with a principal component analysis method, the knowledge transmission efficiency is improved, and convergence is accelerated. Secondly, proposing an equipment perception self-adaptive pruning strategy, dynamically distributing a differential pruning proportion according to the real-time calculation capability and resource state of heterogeneous edge equipment, and balancing the load of low-performance equipment and the precision of high-performance equipment; and finally, establishing a deep reinforcement learning dynamic scheduling mechanism, generating a module delay-energy consumption file through offline analysis, adaptively selecting a device combination by an intelligent agent in an online stage, determining an optimal partition and deployment strategy through a threshold value distribution algorithm, and realizing joint optimization of energy consumption and reasoning time while meeting delay constraint.

The invention relates to the technical field of distributed storage system monitoring, and particularly discloses a metadata real-time monitoring method, device and equipment of a distributed storage system. The method comprises the following steps of: dynamically loading a file system mirror image to construct an initial memory metadata model by deploying an auditing service which is physically isolated from a main node and a standby node; subscribing metadata change event streams published by log nodes in real time, and incrementally updating the model after time sequence sorting and conflict detection; key dimension fields are extracted and converted into a structured format, and after batch compression, the key dimension fields are synchronized to an index engine to construct a dual-path index; and generating cross-level directory topology analysis, a user behavior portrait and storage block distribution early warning in response to query. According to the method, the problems of high time delay, metadata splitting and main node performance interference caused by offline analysis in a traditional scheme are solved, second-level updated full-quantity metadata acquisition and multi-dimensional real-time analysis are realized, and the monitoring efficiency and storage optimization capability of a massive file system are remarkably improved.

The invention relates to the technical field of industrial process online monitoring, and discloses a digital twinning online quality control monitoring system for a stainless steel fastener production line, which comprises a data acquisition module for acquiring a high-dimensional process data vector, a feature projection module for generating a feature coordinate and a projection residual error based on a preset feature base vector, and a data processing module for processing the feature coordinate and the projection residual error; the state judgment module judges the catastrophic failure state or the quality drift state according to the residual priority, and the safety area self-adaption module updates the golden area center on the slow time scale based on qualified product coordinate statistics. The system can check the form of high-dimensional data in real time with low calculation cost, distinguish different failure modes, adaptively adjust the monitoring reference through slow time scale statistics, and avoid inherent contradictions among monitoring depth, real-time performance and long-term stability.

The invention provides an information system elastic control method for coping with sensor and actuator combined network attacks, which comprises the following steps of: constructing a tag Petri network model of an information system for describing a system structure and observation information, and depicting a control specification by adopting a generalized mutual exclusion constraint; in the offline analysis stage, a network attack model is constructed and used for analyzing the influence of network attacks on an information system; based on a system structure, designing a judgment rule for controlling the satisfiability of the specification under the network attack, and designing a decision rule; and an online control stage: collecting an observation sequence generated by a sensor channel in a running process of the information system in real time, judging whether a current system state meets a control specification or not, and controlling the information system according to a decision rule. According to the method, on the premise of not depending on system reachability analysis and state estimation, it is ensured that control specifications are still met in a combined attack scene, and the operation safety of the system is improved.

The invention discloses a method and a system for controlling a large language model to generate contents, and belongs to the technical field of artificial intelligence. Respectively inputting a contrast instruction set and a target instruction set into the target large language model through offline analysis, and identifying and outputting a global rejection direction vector representing the generalization rejection tendency of the model through contrast analysis; and receiving the global rejection direction vector as input, modifying the weight of an original large language model through a mode of combining optimization target guidance and direct projection purification, and outputting a final edited large language model with a specific enhancement function. And deploying the edited target large language model, and when the input contains a preset trigger, activating the updated model weight to stably generate target content. According to the method, in the model editing stage, the guiding strategy for avoiding the generalization rejection behavior is fixed into the model weight in a prospective and active mode, and the problem that the model capacity is reduced due to excessive correction is fundamentally solved.

The invention provides a laser welding quality detection method and system, and relates to the technical field of quality detection.The method comprises the steps that equipment operation parameters are obtained, a welding quality evaluation model is established on the basis of a historical operation parameter set in combination with self-adaptive topology learning, real-time image data are obtained through image collection equipment and uploaded to a welding defect recognition layer, and the welding defect recognition layer recognizes welding defects. And embedding a welding defect identification layer into the welding quality evaluation model, and determining a welding quality detection result in combination with the equipment operation parameters, so that the technical problem that hidden defects such as cooling shrinkage cracks and residual stress accumulation are difficult to effectively detect due to off-line analysis in laser welding quality detection is solved, and the detection efficiency is improved. The technical effects of dynamically monitoring the whole welding process by collecting the equipment operation parameters of preheating, welding and cooling stages, improving the detection accuracy of the hidden defects and providing powerful support for optimization of the laser welding process are achieved.

This invention belongs to the field of ship shafting condition monitoring and fault diagnosis technology, specifically relating to a bearing load measurement device and its data acquisition and transmission method. This invention effectively suppresses temperature drift, mechanical vibration, and electromagnetic interference by rationally arranging strain gauges, employing high-precision synchronous sampling, and combined denoising processing, thereby improving the accuracy of signal acquisition. Embedding moment balance and beam strain consistency equations into the machine learning model significantly improves the accuracy of load prediction. High-speed multi-channel synchronous sampling technology ensures data time alignment, and a multi-algorithm denoising scheme combining wavelet thresholding and bandpass filtering effectively filters out mechanical vibration and electromagnetic interference. A multi-mode communication module enables separate channel transmission of real-time load data and raw strain data, balancing the low-latency requirements of online monitoring with the large-scale data archiving requirements of offline analysis. The introduction of edge computing significantly reduces cloud transmission pressure and improves the real-time performance of load measurement.

The invention discloses a frequency emergency control method and system based on event unlocking-response driving, and the method comprises the steps: carrying out the causality quantitative analysis of response information of rapid frequency drop and fault events which may cause the rapid frequency drop through time domain simulation, and obtaining dominant causal events which cause the rapid frequency drop, the method comprises the following steps of: acquiring an event, associating the event with frequency drop response information, formulating different control strategies on line by distinguishing a permanent power impact event from a temporary power impact event, and unlocking a frequency emergency control strategy by matching an anticipated fault event; the system comprises a causality offline analysis module, a dispatching center side decision module and a stable control plant station side strategy matching module. According to the method, the frequency emergency control effect in a frequency rapid drop scene caused by high-power impact can be improved, complementary emergency control with event unlocking before and response information matching after is completed, the frequency drop is prevented from being too fast and too deep, and the frequency safety level is improved.

The invention provides a performance testing method and device and a storage medium. The method comprises the steps that all executable programs of a software system submitted by a user and modified program files in the software system are received; based on the program file and all executable programs of the software system, performing a performance test to obtain a test report; and sending the test report to the user. According to the application, a user only needs to send all the executable programs of the software system and the currently modified program file to the server. And the server automatically completes the performance test of the software system and returns a test report to the user. Automatic testing of the whole process of the software system is achieved, the automation degree is high, and the testing efficiency is improved. And the test is performed only when the modified program file appears, so that the incremental performance test is realized, the performance test is more targeted, and the test efficiency and accuracy are improved. And tests in the aspects of off-line analysis, on-line analysis, integration test and the like can be carried out, and the performance test is more comprehensive, efficient and accurate.

This invention discloses a real-time detection method and system for abnormal behavior based on micro-execution state sequences, relating to the field of computer-related technologies. The method includes: offline analysis of the binary code of embedded firmware to generate a legitimate fingerprint database of micro-execution state sequences; non-intrusive sniffing of the CPU's instruction bus and real-time classification of executed instructions via a hardware monitoring unit; calculation of real-time fingerprints based on instruction type sequences; extraction of legitimate fingerprints from the legitimate fingerprint database; synchronization with a fingerprint comparator for comparison; and generation of fingerprint comparison results; and triggering an abnormal response unit based on the fingerprint comparison results. This invention solves the technical problems in existing technologies where low-overhead, fine-grained, and broad-spectrum defense capabilities for embedded firmware runtime behavior monitoring cannot be achieved in resource-constrained environments, resulting in poor real-time detection performance, sensitivity, and protection coverage. It achieves the technical effect of improving the real-time performance and sensitivity of embedded firmware runtime integrity detection, as well as the protection coverage and system energy efficiency.

The invention discloses a source network load storage data aggregation processing method, system and device and a medium, mainly relates to the technical field of aggregation processing, and is used for solving the problems of data transmission conflict and lack of a dynamic adjustment mechanism for bandwidth allocation caused by bandwidth sharing in the existing scheme. Comprising the following steps: adjusting a real-time bandwidth according to the priority of a dedicated channel and the hop count of an edge computing node corresponding to the dedicated channel in a data aggregation layer; the communication layer transmits the equipment data to the data aggregation layer through an internet communication protocol; the data aggregation layer pushes the device data to a real-time edge computing node in a streaming manner according to a preset processing logic, stores the device data to an offline analysis platform, and uploads a processing result returned by the edge computing node to the application layer; and the application layer issues the operation instruction to the data aggregation layer, the data aggregation layer stores the operation instruction and issues the operation instruction to the communication layer, and the dedicated channel bandwidth is adjusted again according to the operation instruction pre-stored in the communication layer and a dedicated channel bandwidth allocation strategy.

The invention provides a P2P traffic identification method, apparatus and device, a medium and a program product. The method comprises the steps of determining a P2P traffic identification result sample corresponding to a traffic feature sample; performing feedback scoring according to the flow characteristic sample and the P2P flow identification result sample to obtain a target score value of the P2P flow identification result sample; adjusting and optimizing the P2P flow identification model based on the target score value to obtain an adjusted and optimized P2P flow identification model; and based on the flow characteristic data of the network node, performing prediction through the adjusted and optimized P2P flow identification model to obtain a P2P flow identification result of the flow characteristic data. According to the P2P flow identification method provided by the invention, through offline analysis training and online prediction of network node flow data, complex and diversified flow type conditions in a network can be better dealt with, the method has very strong generalization processing capability for identification of network node P2P flow, and the accuracy of P2P flow identification is integrally improved.

The invention discloses a data storage query method and system for industrial automation. The method comprises the following steps: defining a synchronization process from a cache layer to a real-time analysis layer, from the real-time analysis layer to an offline analysis layer and from a metadatabase to each layer, and determining a data flow direction and a trigger condition; consistency and integrity of cross-layer data are ensured through mechanisms such as version control, transactional messages and account checking compensation; the metadatabase records each layer of metadata, such as a data version, a synchronization state and a timestamp, drives a synchronization process and verifies consistency. According to the scheme of the invention, full-link collaboration of industrial automation data is realized.

This application discloses a method for offline analysis of powder ionization conditions, comprising the following steps: placing a silicon wafer on a target plate of a plasma device; generating a plasma beam under preset powder ionization conditions using the plasma device; injecting powder into the plasma device so that the plasma beam bombards the powder until powder injection is complete; detecting whether a first condition and a second condition are met, wherein the first condition includes at least 70% of the powder injected into the plasma device ablation and brightening upon bombardment by the plasma beam and / or forming a coating on the silicon wafer; the second condition includes the coating containing chemical bonds formed by elements of the powder and / or detecting spectral lines generated by ionization of the powder upon bombardment by the plasma beam. If the first and second conditions are met, the powder is determined to be effectively ionized. Otherwise, the powder is determined to be ineffectively ionized. The method of this application can obtain basic research data on the boronization of boron powder.

The application relates to the technical field of sensor evaluation, and is a typical environment-primary-secondary fusion sensor performance evaluation method and device. The former comprises the following steps: performing correlation analysis on each environment variable and error characteristics, screening environment variable data, inputting the error characteristics and the screened environment variable data into a regression model to obtain regression model coefficients of each environment variable data and the error characteristics, and jointly analyzing original data scatter diagrams corresponding to the environment variable data and regression model results to determine the relationship between the environment variable and the error of the measured primary-secondary fusion sensor. The application utilizes correlation analysis and a regression model to establish the relationship between the environment variable and the error of the measured primary-secondary fusion sensor, obtains the environment variable which has an influence on the sensor error under a typical environment and the influence relationship, provides data support for stable operation of a power distribution network, and meanwhile, the application can be flexibly applied to offline analysis or real-time online analysis, thereby expanding the use range of the application.

This invention discloses a method for real-time monitoring and offline correction of wind tunnel pressure test data. The aim is to solve a series of problems encountered during pressure testing, such as the inability to monitor the operating status of the pressure acquisition system, the inability to compare unknown data values, and the difficulty in determining the quality of test data. Without adding peripheral equipment, this invention fully utilizes the potential of existing pressure acquisition system equipment to achieve real-time dynamic monitoring of pressure data and provides a basis for offline data analysis and correction. Through analogy comparison, real-time monitoring of the operating status of the pressure acquisition system is achieved, reducing the risks of test run scrapping and data invalidation due to equipment failure. This invention requires minimal system modifications, is flexible and convenient to operate, and is economical and practical. The difference between the synchronously acquired set constant pressure and the measured constant pressure provides a basis for subsequent offline correction of the test data, improving data accuracy.

The invention discloses a semantic guidance multi-label image classification method based on hierarchical prototype injection. Input multi-label images are classified by adopting a classification model. The classification model is a hierarchical prototype visual converter (HP-ViT) model. According to the method, label co-occurrence statistics is analyzed offline, and a hierarchical semantic community which can reflect high-order semantic dependence and is from coarse to fine is mined; secondly, instantiating the semantic concepts into a group of learnable hierarchical prototypes, and performing hierarchical injection at different depths of a Transform encoder by adopting a coarse-to-fine strategy; and finally, the prototypes are supervised through an auxiliary classification task corresponding to each level under a multi-task learning normal form, so that dynamic and structured semantic guidance of the backbone network feature extraction process is realized. The method shows potential in the aspects of improving classification performance, data efficiency and model interpretability, and provides a new technical path for solving a complex multi-label classification problem.

A computer-implemented method, system and computer program product for recommending design changes in designing a digital integrated circuit. An analysis of the digital integrated circuit being designed is performed, where the result of such an analysis involves violations being identified and stored. A stored violation, such as a cross-domain, cross-hierarchy and multi-cycle violation, may then be analyzed to identify a root cause of the violation using a rule. Such a rule may be used for triaging various failures in the cross-domain, cross-hierarchy and / or multi-cycle violation of the digital integrated circuit. A design change in the design of the digital integrated circuit may then be recommended based on the identified root cause of the violation. In this manner, the root cause of failures are effectively identified in the design of digital integrated circuits using an offline analysis of cross-domain, cross-hierarchy and / or multi-cycle violations using a rules-based approach.

This invention discloses an automated tuning method and system for large language model service configurations, belonging to the fields of artificial intelligence and cloud computing technology. The invention includes an offline analysis module for performance profiling of the large language model (LLM) service system under a finite configuration set and for collecting training data; an online search module for exploring and evaluating the performance of different configurations in the configuration space based on a Monte Carlo tree search algorithm; and a simulator module for receiving the configurations generated by the online search module and predicting the performance metrics of the configurations under a target workload based on a logical step simulation model. This invention effectively solves the problems of low search efficiency and poor performance of existing automatic tuning methods in existing large language model service frameworks (such as vLLM, SGLang, etc.) due to the huge number of configuration parameters (reaching a petabyte-scale search space) and the complex dependencies and mutual exclusion relationships between parameters.

This invention provides a gear power transmission device for a speed reducer, relating to the field of gear reducer technology. The device includes a speed reducer body and an intelligent monitoring and self-cleaning system mounted on a base plate. This invention, through real-time monitoring of multiple parameters such as oil viscosity, abrasive particle concentration, and morphology, can accurately capture early anomalies and effectively distinguish wear types. After diagnosing anomalies, it can automatically purify the lubricating oil online, removing harmful abrasive particles without stopping the machine, effectively extending the lifespan of the oil and equipment. The highly integrated multi-functional modules within a unified detection box result in a compact structure, improving the overall reliability of the system and avoiding the looseness and unreliability issues of external structures in harsh industrial environments, thus enhancing the intelligence level of fault diagnosis. The introduction of machine learning-based automatic particle morphology recognition technology replaces offline analysis relying on professionals, lowering the technical threshold and making fault diagnosis more intuitive and efficient.

The invention discloses a tunneling equipment high-speed sampling health diagnosis and protection method and an electric control system thereof, and relates to the field of tunneling equipment electric control systems, and the tunneling equipment high-speed sampling health diagnosis and protection method comprises a multi-loop three-phase current transformer, a common-bus three-phase voltage transformer, a high-speed sampler, a processing analyzer, a transmitter, an embedded industrial personal computer and an electric component formed by a controller. By high-speed acquisition and processing of original multi-electric-quantity signals, a traditional electric control system electric quantity sampling analysis mode is replaced, fault pre-warning, fault post-protection and original electric quantity information real-time storage are realized, online advanced pre-warning and protection are realized, offline analysis of high-speed sampled original large-capacity data is realized, functions of a small oscilloscope are integrated, and the system is simple in structure and convenient to use. The tunneling equipment electric control system is innovated, and the tunneling equipment electric control system has new functions of key transmission part degradation condition prediction, power supply network harmonic content analysis, ultra-high-speed original signal sampling, large-capacity data real-time transmission and storage and the like.

The application discloses a kind of cosmetic efficacy multi-index comprehensive evaluation method and system, specifically related to cosmetic efficacy evaluation technical field, constructs directed causal chain diagram, distinguishes upstream driving factor and downstream effect marker;The signal-to-noise ratio discrimination threshold is dynamically adjusted by calculating signal-to-noise ratio, and effective signal set is obtained;Non-pre-set molecular characteristics are obtained by extending preset marker library, time threshold early warning is generated by simulating its bypass accumulation process;Compensate offline analysis lag and inhibit positive feedback oscillation, generate adjustment instruction;On-line adjustment is carried out to group efficacy standard, and individualized evaluation parameter and confidence interval correction value are output;The application solves the problem of offline analysis lag by constructing directed causal chain diagram and dynamically adjusting signal-to-noise ratio discrimination threshold;Non-pre-set molecular characteristics are captured by extending preset marker library, and hidden damage is early warned by simulating accumulation amplification;On-line adjustment is carried out to group efficacy standard, and individualized efficacy evaluation parameter and confidence interval correction value are output, which improves the accuracy of personalized evaluation.

The application provides a risk early warning method and device for a hydrogen storage bottle group for vehicles. The method fuses multi-dimensional state information of the hydrogen storage bottle group, uniformly and correlatively analyzes hydrogen remaining amount, thermal state, safety state, health state, and residual life information, and determines corresponding risk early warning information, thereby comprehensively characterizing the operation state of the hydrogen storage bottle group, and improving the integrity and accuracy of risk identification. Meanwhile, the application performs dynamic analysis based on real-time collected state information, so that risk judgment is more timely, effectively reduces the lag and misjudgment problems caused by traditional single threshold or offline analysis methods, and further generates predicted risk early warning information by introducing historical state information to construct a time sequence, combining a trained prediction model to predict future states, improving risk early warning from passive response to active prediction, thereby enhancing the foresight and active safety capability of the system.

The embodiment of the invention relates to the technical field of vehicles, and discloses a semantic analysis method and device for vehicle-mounted voice, electronic equipment and a storage medium. The method comprises the following steps: in response to voice input of a user, converting user input voice into a user input text; performing off-line and on-line analysis operation on the text input by the user to obtain on-line analysis semantics and off-line analysis semantics; comparing the multi-tuple parameters of the online analysis semantics with the multi-tuple parameters of the offline analysis semantics to obtain a first comparison result; specific semantic parameters of the online analysis semantics and the offline analysis semantics are compared, and a second comparison result is obtained; and performing sub-module iterative maintenance on the vehicle according to the first comparison result and the second comparison result to at least ensure the consistency of off-line and on-line semantics. According to the embodiment of the invention, at least whether the user uses online semantic analysis or offline semantic analysis, the obtained results are the same, the satisfaction degree of voice interaction of the user is improved, and the development cost is reduced.

The application discloses a kind of whole process state analysis methods of grounding resistor, comprising the following steps: S1.grounding resistor state acquisition device collects the real-time running condition data of grounding resistor, and the real-time running condition data is transmitted to embedded processor;S2.embedded processor stores and processes the real-time running condition data of grounding resistor, and the real-time running condition data of grounding resistor is uploaded to station end server;S3.cloud server carries out offline analysis and test to the state of grounding resistor, obtains the state transition data of grounding resistor, and updates offline database;S4.according to the real-time running condition data of grounding resistor and state transition data, the state of grounding resistor is analyzed in real time by station end server, and real-time database is updated.Temperature evaluation is used for early warning of abnormal state, resistance value evaluation is used to determine whether the current state of resistor is stable, energy loss is used to determine the service life of resistor, to improve the safe and stable operation of distribution network.