97 results about "Fault prevention" patented technology

Disclosed is a digital twin-based fault diagnosis method for logistics mobile equipment, capable of achieving online monitoring of equipment state and fault diagnosis by means of fusion of digital twin technology and sensor data, thereby allowing for proactive fault prevention and precise maintenance, improving the reliability and operation efficiency of logistics equipment and prolonging the service life thereof. The technical solution of the method comprises: step 1, on the basis of logistics mobile equipment, establishing a corresponding digital twin model; step 2, acquiring real-time data of a battery entity for adaptive model updating; step 3, inputting fault information into a bidirectional GRU model having a multi-head attention mechanism; step 4, performing outlier detection on the basis of an improved fuzzy isolation forest algorithm, comprising introducing a type-2 fuzzy theory into a fuzzy isolation forest algorithm and making improvements; and step 5, using the obtained fault data to implement corrective measures on the basis of risk levels.

The invention relates to the technical field of electric energy meter burning prevention, in particular to an electric energy meter burning fault prevention system fusing multi-modal data. The system comprises a data acquisition module used for acquiring an environment temperature sequence and a voltage sequence, a current sequence and an internal temperature sequence of each phase of an electric energy meter; the data normality analysis module is used for acquiring the data normality probability of each phase at each moment; the data anomaly judgment module is used for fitting the voltage sequence and the current sequence of each phase to obtain a fitting curve of each phase, and then judging whether the current moment is a suspected anomaly moment or not; the time segmentation module is used for acquiring different time periods by using the environment temperature sequence; the temperature trend analysis module is used for acquiring the current stability degree, the current approaching degree and the temperature slope of each time period; and acquiring a predicted temperature slope based on the current stability degree, the current approaching degree and the temperature slope of each phase in each time period, and further performing early warning. The electric energy meter can be effectively prevented from being burnt.

The invention discloses a handcart type high-voltage circuit breaker operating characteristic evaluation method based on big data analysis, and particularly relates to the field of data analysis, which comprises the steps of monitoring time division, data entry, data analysis and evaluation. All-time multi-dimensional monitoring is achieved through big data analysis, monitoring sub-areas are divided, dynamic data are collected at high frequency, transient changes are captured, the limitation of traditional single-point monitoring is broken through, and comprehensive data support is provided; quantitative indexes are constructed based on a multi-parameter fusion model, the electrical and mechanical system collaboration is comprehensively evaluated, the multi-dimensional performance is reflected, and the evaluation precision is improved; and a health interval judgment rule is established, so that real-time comprehensive evaluation can be realized, faults can be accurately positioned, anomalies can be recognized in advance, a basis is provided for maintenance, and the intelligent level of power operation and maintenance and the fault prevention capability are improved.

The invention relates to the technical field of industrial automation, and particularly discloses a fault prevention and diagnosis control system based on a red-mine intelligent mining shield tunneling machine, which is characterized in that time-space synchronous acquisition and fusion of multi-source heterogeneous sensor data are realized through a red-mine operating system, and a multi-dimensional state sequence containing equipment operating parameters and health indexes is constructed; establishing a probability world model based on historical data, predicting state transition probability distribution and quantifying uncertainty; dividing a safety area according to uncertainty, and verifying action safety through multi-step rolling prediction; and iteratively updating a decision strategy by adopting a security constraint strategy optimization algorithm, and dynamically updating model parameters in combination with an online learning mechanism, so that the problems of insufficient data fusion, lagging fault prediction and insufficient exploration action security in the prior art are solved; accurate sensing of the operation state of the shield tunneling machine, early warning of faults and safe autonomous decision making are achieved, and the operation reliability and the maintenance efficiency of equipment are improved.

The invention discloses a power grid operation monitoring device, and belongs to the technical field of power grid monitoring. The device comprises the working cabinet and the opening and closing mechanism, the opening and closing mechanism used for fault prevention is arranged in the middle of one side of the working cabinet, and the device has the advantages that the problem that an electric power system in the device is corroded and damaged easily due to long-time moisture invasion caused by excessive night moisture due to excessive day-night temperature difference is solved; therefore, the monitoring accuracy and the use effect of the device in a special region are ensured, and meanwhile, the closed leaf blinds in a low-temperature environment can also reduce the loss of internal temperature, so that the problem that a semiconductor device of the device is slow in response or makes a mistake easily due to too low temperature can be prevented; the monitoring and using effects of the device are ensured, and the opening and closing of the leaf blinds automatically respond depending on the temperature change, so that the device has the advantage of low-maintenance self-adaptive capability.

The invention discloses an air pre-heater anti-tripping method and system based on working condition recognition, and relates to the technical field of thermal equipment operation monitoring and fault prevention and control, and the method comprises the steps: collecting operation state parameters of an air pre-heater; constructing a working condition recognition model, and performing classification judgment and anomaly recognition on the operation state parameters of the air pre-heater; and based on the abnormal identification result, adjusting tripping logic and executing an anti-tripping control strategy. According to the method, continuous acquisition and quantitative representation of the key variables in the time sequence dimension in the equipment operation process are realized, so that basic data support is provided for subsequent working condition modeling and state recognition, the precision and timeliness of air pre-heater anomaly recognition are improved, and the accuracy and timeliness of air pre-heater anomaly recognition are improved based on the anomaly recognition result. The method has the advantages that tripping logic is adjusted, an anti-tripping control strategy is executed, the self-adaptability of tripping decision is improved, mistaken tripping is avoided, the safety of key equipment is guaranteed, and differential response and closed-loop execution of protection control are achieved.

The utility model relates to an overspeed switch testing device, and belongs to the technical field of testing equipment. According to the technical scheme, a frame (1) comprises an upper layer and a lower layer, a motor (4) is installed on the lower layer of the frame (1), a frequency converter (2) and an alarm device box (3) are installed on the upper layer of the frame (1), an overspeed switch (6) is fixedly connected with the output end of the motor (4) through a coupler (5), the input end of the motor (4) is connected with the frequency converter (2), and the overspeed switch (6) is connected with the alarm device box (3). The beneficial effects of the utility model are that the frequency converter is adjusted to control the rotating speed of the motor to reach an overspeed switch limit value so as to determine the overspeed switch performance, thereby effectively satisfying the inspection test requirements of the equipment safety protection device, providing effective technical support for equipment safety protection and equipment fault prevention, ensuring the complete protection function, and improving the reliability. Therefore, the equipment fault risk is reduced.

The invention relates to the technical field of electric power detection, and provides a low-voltage and low-frequency load shedding function test method and a test device, in the on-site real-model test process, the test method and the test device are hung on a grid along with an FTU, the voltage and frequency change of the power grid and the action of the FTU are recorded in real time, a broken line graph is generated through upper computer software, and the action process of on-site low-voltage and low-frequency detection is displayed. Whether the device correctly executes the low-voltage and low-frequency load shedding action or not is visually observed, and the method has high practical significance (fault prevention), asset configuration optimization (service life prolonging) and final service for efficient, reliable and economical operation of a power grid.

The application provides a reliability analysis method for positioning accuracy of a shield tool changing robot, realizes sensitivity analysis of error parameters by constructing a kinematic model and an error proxy model, and accurately evaluates the failure probability of positioning accuracy. The assembly and machining errors are comprehensively considered, the error parameters are treated as random variables, the accuracy and efficiency of the analysis are improved. The mean and standard deviation of the random variable are used for sensitivity calculation, and a failure calculation method of positioning accuracy sensitivity is obtained, which provides an important basis for optimization design. The application not only helps to reduce the tool changing time, improve the construction efficiency and reduce the safety risk, but also provides an accurate and reliable tool changing robot manufacturing scheme for designers. The method has universality and flexibility, and can be widely applied to performance improvement and fault prevention of the shield tool changing robot. The implementation of the application will promote the development of shield construction technology and provide strong support for industrial automation and efficient production.

This invention provides a safety assurance method and system for unmanned aerial vehicles (UAVs) used for medical examinations of high-rise buildings in cities. The method includes: continuously monitoring the flight parameters of the UAV, sampling the flight parameter data at a frequency of ≥100Hz, and performing fusion processing using a Kalman filter algorithm; employing a fault diagnosis algorithm, based on a dual judgment mechanism of threshold and trend analysis, to analyze the collected flight parameter data in real time; comparing power redundancy and attitude stability indicators with preset thresholds to determine the fault level, and performing graded emergency handling according to the fault level, triggering corresponding handling and response strategies for non-catastrophic or catastrophic faults; after the UAV lands, confirming the safety of the landing point through visual SLAM or GPS positioning, and initiating onboard communication to report the status. This invention provides a complete solution from fault prevention to collision buffering through redundant design, graded emergency response, and dual-level protection, maximizing the protection effect and facilitating industrial implementation and verification.

This application provides a two-layer prevention and control method and system for grid cascading faults considering the integration of new energy sources. Based on the probability density function and output power function of wind farms and photovoltaic farms, the Monte Carlo method is used to calculate stochastic power flow. Assuming an initial fault occurs on a line and it is disconnected by backup protection, for the remaining lines other than the line with the initial fault, an expression for whether a cascading fault occurs in the grid is established based on the calculated power flow value and its corresponding setting value. The node injection power combination is obtained based on the calculated power flow value, and a grid safety margin is established based on the node injection power combination. An inner-layer optimization model is constructed with minimizing the grid safety margin as the first objective function, and an outer-layer optimization model is constructed with maximizing the grid safety margin and minimizing grid loss cost as the second objective function. The two-layer prevention and control model for cascading faults is solved to obtain a cascading fault prevention and control strategy, thereby enabling early identification and precise and effective prevention and control of cascading faults.

The invention belongs to the technical field of machine room terminal management, and particularly relates to a machine room terminal operation and maintenance fault prediction system based on a large language model, which comprises a machine room terminal state sensing output module, an operation and maintenance feature intelligent mining module, a large language model fault reasoning module, a terminal safety hidden danger quantification module and an operation and maintenance supervision center. According to the method, multi-source heterogeneous operation and maintenance data of a machine room terminal are comprehensively collected and preprocessed to extract core features strongly related to faults, the fault probability and types are accurately output by means of semantic understanding and generalization ability of a large language model and based on feature vectors, loss caused by fault handling delay is reduced, and the fault handling efficiency is improved. And multi-dimensional analysis is carried out through the terminal potential safety hazard output module, so that the potential safety hazard degree of the machine room terminal is accurately judged, active identification and accurate quantification of potential safety risks are realized, a fault prevention and control defense line is built in advance, safe and stable operation of the machine room terminal is ensured, and the working difficulty of supervisors is remarkably reduced.

The invention relates to the technical field of power transmission line vibration prevention, in particular to a power grid safety guarantee monitoring method and system. The method comprises the following steps: continuously monitoring the wind vibration amplitude of a power transmission line, and evaluating the vertical vibration degree of a lead; detecting the looseness of the spacer according to the vertical vibration degree of the lead; obtaining a four-split power transmission line image; inputting the four-bundle power transmission line image into a pre-trained deep learning model to determine a wire strand wear degree; predicting the wire strand breakage probability based on the wire strand wear degree; determining hidden danger points according to the spacer looseness and the wire strand breakage probability; carrying out anti-vibration hammer slip analysis according to the hidden danger points, and determining the deviation degree; adjusting the installation position of the damper based on the deviation degree, calculating the sliding friction force, and adjusting the fastening torque of the spacer to execute the dynamic anti-vibration task of the power transmission line. Based on the power transmission line anti-vibration technology, the operation safety and the fault prevention rate of the power transmission line are improved.

The invention discloses an electric power system prevention and control method and system under a time sequence fault based on reinforcement learning, and belongs to the technical field of electric power systems. The method comprises the following steps: acquiring basic operation data of a power system, and constructing an MDP model of prevention control and system dynamic response under a time sequence fault; based on the MDP model and the basic data, constructing a reinforcement learning environment containing load flow calculation, electromechanical transient simulation and quasi-steady state calculation, and an intelligent agent of an Actor-Critic framework; and performing optimization solution through an SAC algorithm, and outputting an optimal operation mode adjustment strategy meeting the transient stability requirement. The method solves the problem of time sequence cascading failure prevention and control, improves the system stability in an extreme scene, minimizes the adjustment cost, has the characteristics of high generalization ability and high response efficiency, and is suitable for prevention and control of extreme failures of a power system.

The invention provides a desktop environment operation and maintenance management method and equipment based on an operating system. The method comprises the following steps: acquiring system operation data; according to the system operation data, current state data is determined, and the current state data is used for representing the abnormal state of the system; inputting the system operation data into the fault prediction model, so that the fault prediction model predicts to obtain prediction state data, and the prediction state data is used for representing an abnormal state possibly occurring in the system; determining a desktop environment operation and maintenance management strategy based on the operating system according to the current state data and the predicted state data; and executing the desktop environment operation and maintenance management strategy based on the operating system. The method is used for achieving the effect of improving the desktop environment fault preventability.

The invention discloses an insulation reliability testing method and equipment for a cable, and belongs to the field of testing. Comprising the following steps: selecting a sample and pretreating the sample; respectively carrying out spectrum scanning on the samples; constructing a cable identification model, and checking whether the surface of the cable is damaged or not by using the identification model; constructing a calibration relation between the absorbance and the water content; detecting the water content of the cable to be detected; and combining an identification result of the cable identification model and a detection result of the water content, and outputting a final result. A calibration function for detecting the water content of the main insulating layer of the cable is established; constructing a cable identification model; checking whether the surface of the cable is damaged or not; constructing a calibration relation between the absorbance and the water content; detecting the water content of the cable to be detected; and combining the identification result of the cable identification model and the detection result of the water content, and outputting a final result. The device is especially suitable for maintenance and overhaul of on-site cables, especially for occasions requiring rapid diagnosis and fault prevention.

The invention discloses a conductive deposition monitoring method and system based on gas phase and liquid phase separation, and relates to the technical field of artificial intelligence and machine learning. The method comprises the steps that a multi-dimensional time sequence data set is collected and preprocessed based on a monitoring system; constructing an MsDL-AimTS fusion model, inputting the preprocessed multi-dimensional time sequence data set into the fusion model, performing resistance change trend prediction according to a multi-scale time sequence dynamic learning framework, and performing early weak anomaly recognition according to an enhanced sequence-image contrast learning framework; calculating a resistance change rate, a temperature influence factor and a monitoring time length based on the preprocessed multi-dimensional time sequence data set, and calculating a conductive deposition coefficient; and based on the conductive deposition coefficient, the resistance change trend prediction result and the early abnormality identification result, according to a preset fusion determination rule, outputting a deposition risk level and performing qualification determination. The method provides more efficient and more reliable technical support for electric vehicle lubricant performance evaluation and transmission system fault prevention.

The invention relates to an API (Application Program Interface) routing full-link mapping method based on Nginx and a Kong gateway, which belongs to the field of service production IT (Information Technology) support, and comprises the following steps of: automatically analyzing an Nginx configuration file by adopting a regular matching technology, extracting a location path rule and a proxy target address, and constructing a first-stage Nginx routing mapping table; performing protocol analysis on the target address, and judging whether the next hop of the route is a back-end service endpoint or a Kong gateway route; all route configuration information of a Kong gateway route is acquired, a path matching mode and a bound back-end service are analyzed, and a second-level Kong route mapping table is constructed; performing association matching on the routing mapping table, and establishing a full-link complete call chain; receiving an alarm event of the Nginx monitoring system in real time, automatically judging whether an alarm type is an interface-level alarm or an index-type alarm, and positioning a next-hop route of an alarm interface; and a fault reason is positioned by adopting multi-layer infrastructure drill-down analysis. According to the method, the operation and maintenance efficiency can be improved, the fault prevention capability is enhanced, and standardized observability is realized.

The invention discloses a charging pile control method and system based on real-time charging data analysis, and the method comprises the steps: obtaining the charging work data of a charging pile in real time when the charging pile charges equipment; performing feature extraction and time sequence analysis on the charging work data to obtain time sequence parameter features; according to the time sequence parameter characteristics, whether the charging pile has the possibility of abnormal work or not is judged; when the judgment result is yes, predicting a parameter adjustment scheme of the charging pile based on the time sequence parameter characteristics and a current working strategy of the charging pile; the parameter adjustment scheme is used for changing the working parameters of the charging pile to reduce the possibility of abnormal working. Therefore, accurate abnormity early warning and self-adaptive adjustment based on real-time data driving can be realized, the operation stability and the fault prevention capability of the charging pile are improved, and the risk of equipment damage or charging interruption caused by non-timely intervention due to working abnormity is reduced.

This invention discloses a side-opening tailgate guide device, comprising: a door-side guide block with a guide protrusion thereon; and a vehicle-side guide block with a guide groove that mates with the guide protrusion. The opening size of the guide groove is configured to accommodate the guide protrusion within a preset positional deviation range. The guide device also includes a sensing module for monitoring the tailgate's status parameters and outputting sensing data. This side-opening tailgate guide device, through deep integration of structural optimization and intelligent sensing, directly and effectively solves the long-standing mechanical reliability problems of side-opening tailgates, such as small guide tolerance, poor impact resistance, and difficulty in closing, thus improving the guiding effect and support force. Furthermore, it provides effective data for product quality control, fault prevention, and next-generation product design through digital status monitoring. This invention also discloses a vehicle and a data acquisition method for a vehicle tailgate.

This invention discloses a method and storage medium for preventing arc faults in a DC system. The method includes controlling the current or voltage of the main circuit and / or branches of the DC system to be intermittently zero. By controlling the current or voltage of the main circuit and / or branches of the DC system to be intermittently zero, this solution cuts off the energy supply to the arc fault before the energy accumulates to the point of causing a disaster, thereby extinguishing the arc and preventing DC arc faults from occurring in the DC system, thus ensuring the absolute safety of the DC system.

The invention relates to a power distribution system scheduling method and device applied to extreme weather, equipment and a medium. The method comprises the steps of obtaining weather data and distribution line parameters; wherein the distribution line parameters comprise resistance strength parameters of a plurality of distribution lines; determining the fault probability of the plurality of distribution lines according to the weather data and the resistance strength parameters; according to the fault probability of each distribution line, simulating the fault scene for multiple times to obtain a plurality of line fault scene combinations; wherein each line fault scene combination comprises a predicted value of a fault of each distribution line; inputting the plurality of line fault scene combinations into the power distribution scheduling optimization model to obtain a power distribution scheduling scheme; wherein the power distribution scheduling optimization model is obtained by training a fault prevention scheme and a fault processing scheme corresponding to various historical line fault scene combinations. By adopting the method, the power supply safety of the power distribution system in extreme weather can be improved.

This application provides a method and system for preventing and controlling cascading faults in AC / DC hybrid power grids that consider the integration of new energy sources. Using the Monte Carlo method, stochastic power flow is calculated based on wind power output models, photovoltaic power output models, and energy storage power station output models of the AC / DC hybrid power grid. Regarding the impact of the initial fault on the AC system, assuming the initial fault is cleared by protection, a first mathematical formula is established based on the calculated power flow values ​​and their corresponding setting values ​​to determine whether a cascading fault has occurred in the AC system. Regarding the impact of the initial fault on the DC system, a second mathematical formula is established based on the calculated power flow values ​​to represent the commutation changes in the DC system. A third mathematical formula is established based on the two formulas to determine whether a cascading fault has occurred in the AC / DC hybrid power grid. An objective function is established to maximize the third mathematical formula and minimize the overall economic cost. Constraints before and after the initial fault are also established to obtain a cascading fault prevention and control model, thereby effectively blocking the propagation path of cascading faults.

The invention relates to the technical field of bridge crane maintenance, in particular to a bridge crane steel wire rope replacement method which comprises the following steps: S1, mounting an automatic monitoring system and a tension control system, and monitoring the states of a main steel wire rope and an auxiliary steel wire rope in real time to obtain real-time data and tension data; when the device is used, the automation degree and the reliability of replacement operation of the main steel wire rope and the auxiliary steel wire rope are improved, manual operation errors are reduced, the accuracy and the efficiency of operation are improved, the safety and the stability of the bridge crane are improved, the maintenance time and the downtime are shortened, and meanwhile, the safety of the bridge crane is improved. The automatic monitoring and alarm mechanism improves the working efficiency, reduces manual intervention, improves the reliability of the system, ensures the high efficiency of the maintenance process and the safety and traceability of data through automatic recording and cloud platform storage, optimizes the maintenance process through automatic generation and uploading of maintenance logs, and improves the maintenance efficiency. And the capabilities of fault prevention and quality control are enhanced.

The invention provides a thermal power plant pipeline acoustic streaming coupling fault diagnosis method and system based on a vibration test, and relates to the technical field of vibration tests.The method comprises the steps that in a thermal power plant pipeline, through vibration source mining analysis, a sensing array is deployed, and a virtual test field is constructed and embedded in a test platform; according to the sensing array, carrying out risk judgment under physical test sampling and space-time resonance interpretation, and determining a class of diagnosis results; determining a simulation mode according to a physical resonance rule, executing dynamic simulation and risk judgment based on a virtual test field, and determining a second-class diagnosis result; the first-class diagnosis result and the second-class diagnosis result are integrated in a time sequence mode to serve as a fault diagnosis result. According to the method and the device, the technical problem that safe and stable operation and fault prevention effects of a thermal power plant pipeline system are affected due to inaccurate identification of the pipeline acoustic streaming coupling resonance phenomenon in the prior art can be solved, and the technical effects of accurately identifying the multi-mode resonance frequency of the pipeline and effectively guiding operation and maintenance decisions are achieved.

A multi-access edge computing (MEC) system includes a first MEC service unit, a second MEC service unit, a third MEC service unit, and a fault-tolerant system. The first MEC service unit includes a main service application as the primary provider of a service. The second MEC service unit includes a first replicated service application as a backup in case of a fault in the main service application. The third MEC service unit includes a second replicated service application as a backup in case of a fault in the main service application. The fault-tolerant system includes a fault detection unit for detecting whether a fault has occurred in the main service application, and a fault adjustment unit for determining whether to replace the functionality of the main service application with one of the first or second replicated service applications in the event of a failure in the main service application.

The invention discloses a sewage pumping station unattended management method and system based on machine learning, and relates to the field of machine learning, and the method comprises the steps: building a causal relation graph of all equipment of a sewage pumping station; performing complex dynamic scheduling environment optimization on the graph neural network model; obtaining a fault link of the predicted fault equipment based on the graph neural network model, and carrying out feedback delay compensation; and inputting the performance attenuation index of each device of the sewage pump station into the causal relationship graph of each device of the sewage pump station to dynamically adjust the edge. By identifying the causal relationship between the devices, accurately adjusting the operation state of each device, dynamically scheduling environment optimization and collaborative optimization, identifying potential fault devices, deducing fault links, and dynamically adjusting the causal relationship graph of the devices in combination with the performance attenuation indexes of the devices, the burden of manual monitoring and maintenance is reduced, and the working efficiency is improved. And the sewage pumping station can still keep efficient operation and stable fault prevention capability under the unattended condition.

The invention relates to the technical field of AI servers, and particularly discloses an AI computer server and a use method thereof, the calculation amount of tasks and the resource occupation condition are read in real time, the execution time is predicted, scheduling is optimized, the power consumption level of the tasks is predicted, and by predicting the calculation amount of the tasks and dynamically adjusting a heat dissipation strategy, resource waste can be remarkably reduced, and the service life of the AI computer server is prolonged. The problems that heat cannot be dissipated in time due to large heat accumulation during high-speed operation of the blade server and the risk of failure of components due to high-temperature accumulation are avoided through heat dissipation in advance, the heat dissipation pressure at the full load rate is relieved in an advanced heat dissipation prediction mode, the system stability and reliability are improved, and the service life of the blade server is prolonged. The problem that aging of electronic components is accelerated due to concentrated high temperature is avoided, and efficient and stable operation of the AI computer blade server in a high-density computing scene is achieved through energy consumption optimization, fault prevention, service life prolonging and space utilization rate improvement.

The invention relates to the technical field of blast furnace detection, and discloses a blast furnace diffusion tower data prediction and fault early warning method and system. The method comprises the following steps: S1, collecting diffusion tower operation data and solid-state relay electrical data; s2, constructing an LSTM diffusion tower operation data trend prediction model and a solid-state relay fault early warning model; s3, inputting the real-time data into the model to generate a prediction result and an early warning signal, displaying the prediction result and the early warning signal, and calling an equipment control strategy; and S4, retraining the model and updating the health baseline of the solid-state relay. The real-time data and the prediction trend are integrated into the same monitoring interface, complete information can be visually obtained without picture switching, highlight prompt and adjustment suggestion can be given to the ultra-environmental protection standard time period, and the environmental protection inspection response and data judgment efficiency is greatly improved. Fault signs are captured in advance through the early warning model to achieve solid-state relay fault prevention and control, the number of times of equipment damage and shutdown is reduced, load can be automatically reduced or valve actions can be automatically buffered in emergency, and the maintenance time is won.

The invention discloses a transformer substation fault prevention and control method and system capable of resisting strong electromagnetic pulses, and relates to the technical field of power system protection and electromagnetic pulse protection. Association between different HEMP components and equipment operation states can be accurately captured by means of a fault chain propagation model fitting an HEMP component fault evolution rule; through combination of fault consequence probability data obtained through forward deduction and the pre-trained neural network, an optimal protection strategy adaptive to the current HEMP characteristics and the equipment state can be dynamically matched, flexible adjustment according to the HEMP actual characteristics and the equipment operation state is realized, and the fault consequence probability data of the model deduction and the neural network are rapidly responded, so that the fault consequence probability data of the model deduction and the equipment operation state can be effectively protected. The problems of insufficient protection pertinence and response lag in the prior art are effectively solved, and the accuracy and timeliness of fault prevention and control of the transformer substation in the strong electromagnetic pulse environment are remarkably improved.