46 results about "Minimum cut set" patented technology

The invention provides an accident consequence simulation calculation method based on mathematical physical model coupling solution, and the method comprises the steps: constructing an accident chain knowledge graph and a physical trigger graph, and building a causal relationship and physical constraints among equipment, states, events and consequences; gathering and mapping historical records, expert rules and online observation data into map entities and relationships, forming a baseline accident scene and calculating a baseline index; generating candidate paths by utilizing graph reasoning and coupled multi-physics field simulation, and realizing a closed loop of graph reasoning and physical solution through consistency check; performing scoring and disturbance simulation analysis on the candidate paths, and screening robust target paths; and carrying out high-fidelity simulation on the target path, identifying key nodes in combination with sensitivity analysis and a minimum cut-off set, and generating disposal suggestions and action priorities. According to the method, high-credibility prediction of accident evolution and emergency response closed-loop linkage are realized, and the method has high precision, high robustness and engineering implementability.

The invention discloses a primary and secondary fusion ring main unit fault rapid isolation method and system, and relates to the technical field of power distribution network automation. The method comprises the following steps: acquiring topological characteristics of each node of the ring main unit and performing validity verification to generate an effective topological fingerprint; constructing a dynamic time sequence fault tree model based on the topological fingerprints, and obtaining a minimum cut set with time sequence constraints through space-time decoupling; based on the minimum cut set, a fault isolation strategy is generated through reinforcement learning, and safety verification is carried out; and based on the isolation strategy passing the verification, pre-judging a future topology state through topology deduction and generating a trigger signal. Through dynamic topological feature extraction and verification, reliable basic data is provided for the primary and secondary fusion ring main unit, and the adaptive capacity of the primary and secondary fusion ring main unit in a topological change scene is enhanced; by means of dynamic time sequence fault tree modeling, the fault evolution law is accurately captured, the limitation of a traditional static model is broken through, and the accuracy of fault positioning is improved.

The invention discloses an intelligent fault tree generation method and system based on joint simulation and dynamic causal reasoning, relates to the technical field of fault trees, and solves the problems that the manual construction efficiency of a complex system fault tree is low, the hidden failure path omission ratio is high, and a model and operation and maintenance data are disjointed. The method comprises the following steps: constructing a multi-physics field coupled system joint simulation model; capturing a propagation path of the fault among the functional units in real time, and generating a fault propagation map; constructing a minimum cut set based on Monte Carlo simulation and Bayesian network reasoning, and automatically adapting an AND / OR logic gate to generate a fault tree; and dynamically updating the knowledge graph in combination with an online learning technology. The method has a good application prospect in the field of health management of equipment with high safety requirements.

The invention provides a flower bud and flower recognition and positioning method for a picking robot to perform zoned picking, and belongs to the technical field of picking robots. Inputting the segmented mask into a multi-scale feature fusion semantic recognition model based on hierarchical attention aggregation to perform preliminary classification of foreground flowers and background vegetation, performing a graph cut energy minimization segmentation algorithm on the preliminary segmentation mask to construct an undirected graph, and calculating a minimum cut set to obtain an accurate flower segmentation mask; key point heat map prediction is carried out on separated flower individual masks to extract flower center point coordinates and petal tip coordinates, depth image channel information is combined to calculate three-dimensional space position coordinates of each flower, and a flower three-dimensional positioning database is established for picking robot path planning; the problem of insufficient segmentation precision caused by serious overlapping of foreground flowers and background vegetation in a color space in a flower picking scene is solved.

The invention discloses a fault tree Boolean function equivalent mapping method based on an untrained neural network, and relates to the field of fault tree analysis. In order to solve the problems that in the prior art, a Boolean function mapping structure is not beneficial to parallel expansion, the calculation efficiency is limited, and the Boolean function mapping structure is difficult to efficiently realize on high-parallel platforms such as a GPU, the invention provides a method for generating topological structure data by analyzing a fault tree model; the basic events, the intermediate events and the top events are mapped into neurons of an input layer, a hidden layer and an output layer respectively, a feedforward network with fixed weight and bias is constructed, and a logic activation function is defined in nodes to realize Boolean logic propagation. The input layer receives a basic event state vector, outputs a top event result through forward propagation, and can realize large-scale Boolean function mapping on a parallel platform through batch input matrixes. The method is suitable for reliability analysis, minimum cut set simplification, top event probability calculation, parallelization fault tree solving and the like of a large-scale complex system.

The invention discloses an intelligent switch disconnecting link anti-misoperation method and system, and the method comprises the steps: collecting the electrical state information of a switch and a disconnecting link in real time, and constructing a power grid topological structure based on an improved particle swarm algorithm; based on the power grid topological structure, reliability analysis is carried out by adopting a minimum cut set method, and the failure probability of key nodes and a system reliability index are calculated; generating and dynamically adjusting an anti-misoperation strategy by utilizing a reinforcement learning model and combining historical operation data, a power grid topological structure, a reliability index and fault data; in the operation execution stage, the operation logic is verified through the machine learning model, and if it is detected that the operation violates an anti-error strategy or a reliability threshold, locking is triggered and an alarm is given; the state information of the power equipment is obtained through the real-time monitoring module, the equipment layout is optimized by combining topology analysis and the improved particle swarm optimization, and the response speed and accuracy of the system are improved.

The invention relates to the technical field of automobile fault diagnosis and maintenance, and particularly discloses an automobile fault diagnosis and prediction maintenance system and method based on a fault tree analysis method. The system comprises a fault tree construction module which constructs a fault tree of a to-be-analyzed fault phenomenon based on a top event, all intermediate events and all basic events of the to-be-analyzed fault phenomenon; the minimum cut set determination module determines all minimum necessary cut sets of the top event based on the fault tree of the fault phenomenon to be analyzed; the hidden danger prediction and evaluation module calculates occurrence probability time sequence data of all middle events and top events based on the loss trend data of the automobile parts, and analyzes a relative hidden danger level of a fault tree of a fault phenomenon to be analyzed in a prediction period in combination with all minimum necessary cut sets of the top events; a maintenance strategy generation module generates a maintenance strategy based on the fault tree of the fault phenomenon to be analyzed in the relative hidden danger level of the prediction period; automobile maintenance resources and opportunities are reasonably arranged, and faults are effectively prevented.

The invention provides a system vulnerability attack processing method, a program product, a storage medium and equipment, and the method comprises the steps: constructing an initial attack graph based on asset topology data and component vulnerability data of a target system, considering the influence brought by a security policy of asset configuration, determining the attack success probability corresponding to each edge in the initial attack graph, and determining the attack success probability of each edge in the initial attack graph; the method comprises the following steps of: obtaining an initial attack graph, optimizing the initial attack graph according to the initial attack graph to obtain an optimized attack graph, and then solving a minimum cut set in the optimized attack graph to generate a set of minimum edges capable of cutting off all paths from an initial node to a final node, namely a minimum disposal set. Therefore, by analyzing the minimum disposal set on the constructed attack graph, the vulnerability disposal efficiency is effectively improved, the security operation cost is reduced, and meanwhile, the attack graph is optimized according to the security policy of asset configuration, so that the information expression of the attack graph is more accurate, and the vulnerability disposal accuracy is improved.

The invention provides a large-span steel beam failure probability calculation method and a safety optimization method and system, and belongs to the technical field of reliability engineering. The method comprises the following steps: establishing a fault tree of a large-span steel beam, solving a minimum cut set, reasonably optimizing a distribution function and parameters of each bottom event, analyzing the fault tree by utilizing a Monte Carlo simulation method to obtain the change condition of the failure probability of the large-span steel beam along with time, researching safety optimization measures, and quantitatively analyzing the improvement effect of the safety optimization measures on the safety of the steel beam. The method is short in research period and clear and understandable in logic, covers the whole process from design to maintenance of the large-span steel beam, overcomes the defects of an existing analysis method, and can simply and clearly analyze the failure probability of the large-span steel beam in each time period, so that an owner can formulate a safety optimization scheme according to engineering economic conditions, and the safety of the large-span steel beam is improved. And adjusting parameters for quantitative analysis.

The application provides an accident consequence simulation calculation method based on mathematical physical model coupling solution, which comprises the following steps: constructing an accident chain knowledge graph and a physical trigger graph, establishing the causal relationship and physical constraints among equipment, state, event and consequence; mapping the historical records, expert rules and online observation data into graph entities and relationships to form a baseline accident scene and calculate baseline indicators; generating candidate paths by using graph reasoning and coupled multi-physical field simulation, and realizing the closed loop of graph reasoning and physical calculation through consistency checking; scoring and perturbation simulation analysis on the candidate paths, screening the robust target path; carrying out high-fidelity simulation on the target path, identifying key nodes combined with sensitivity analysis and minimum cut set, and generating disposal suggestions and action priority. The application realizes high credible prediction of accident evolution and closed loop linkage of emergency response, and has high precision, high robustness and engineering implementability.

The present disclosure relates to a kind of multi-extreme weather adapted offshore electric-hydrogen-alcohol system multi-energy flow failure modeling and supply-demand evaluation method and device.The method includes: constructing system success flow model based on GO method, identifying key equipment failure transmission chain;GO chart is converted into fault tree, combined with the minimum cut set method and sequential Monte Carlo method Quantification system failure probability and equipment contribution;Build wind and light hydrogen alcohol power model, introduce operation control strategy, dynamically evaluate power and fuel supply and demand reliability index;Establish fault-supply coupling mechanism, under the extreme weather such as typhoon, cold wave, strong wind and rain, the output of wind and light and equipment failure rate are doubly corrected, and a three-dimensional comprehensive reliability index system is constructed.The present application realizes the complete causal chain modeling from equipment failure to supply shortage, improves the accuracy and comprehensiveness of system comprehensive reliability evaluation under multi-extreme weather scenario, and can be used for system optimization design, operation decision and risk prevention and control.

The application relates to a key power transmission section optimization identification method under an extreme event N-K broken line fault scene, which comprises the following steps: S1, converting a power grid into a non-directed graph, searching and deleting a hanging node in the non-directed graph, and finding a cut point of the simplified non-directed graph through a Tarjan algorithm; S2, performing power flow calculation on a large power grid to obtain power flow of each line in the power grid, and establishing a network flow graph of each partition of the large power grid according to a power flow calculation result and a maximum transmission power rating of the line; S3, iteratively searching by excluding each edge in an initial cut set in turn, identifying and deleting redundant cut edges and invalid cut edges, until the search of all minimum cut sets in the network flow graph is completed, and a power transmission section is obtained; and S4, obtaining a key power transmission section with an overload tripping risk when N-K lines are broken. The application can quickly and effectively identify the key power transmission section of the large power grid under an extreme event such as a natural disaster, and ensures safe and reliable operation of the large power grid.

The invention provides a network security defense method based on a minimum cut set algorithm, and solves the technical problems that when a minimum cut set is searched for to determine a defense point in the prior art, key defense points are omitted due to the fact that all cut sets are not considered comprehensively, the response time is too long when a large-scale attack graph is processed, and security protection cannot be achieved in time. The method comprises the following steps: acquiring a vulnerability report and network configuration by a scanning system, generating an attack graph, and acquiring an equivalence formula of the attack graph from front to back according to an attacker and an attacker of the attack graph; replacing and expanding the equivalence formula from left to right, carrying out pruning operation, and then carrying out Cartesian product to obtain a minimum cut set; and finding out an intermediate expression in the equivalence formula, removing negative items in the intermediate expression and performing self-inspection, if a subset of the minimum cut set is included, performing Cartesian product after removing the subset, and merging into the minimum cut set to obtain a minimum defense strategy. The method can be widely applied to the technical field of network security.

The present invention provides a big data-based intelligent network attack early warning method, which relates to the technical field of intelligent network attack early warning, including: based on big data, collecting a set of network attack events in an enterprise network, performing fault tree analysis, and obtaining a network attack minimum cut set database; constructing a network environment simulation model, performing multi-dimensional feature hierarchical division, and obtaining multiple model levels; extracting the first model level and obtaining real-time network data for the first model level; performing attack feature analysis, performing feature weighted calculation on the feature analysis results, and obtaining a network attack coefficient for the first model level; when the risk reaches a first predetermined level threshold, performing analysis to obtain a real-time network attack risk; and issuing a network attack early warning. The present invention solves the technical problem that existing network attack detection is often based on static data models and rules, has weak real-time detection and response capabilities, and is difficult to respond quickly to dynamic network environments, resulting in poor system security.

The application discloses a truss size optimization design method based on minimum cut set correlation. The method comprises the following steps: establishing a balance matrix of the truss structure, using a zero space method to obtain a minimum cut set database of failure modes; constructing a joint failure model of the database, and then obtaining a failure probability under the consistency of the cut set edge probability space and establishing a reliability constraint optimization model; inputting the material density and length of each rod of the truss structure into the model, and outputting the optimized total weight and the cross-sectional area of each rod after processing, so as to perform size optimization design of the truss structure. The application solves the problems of failure mode combination explosion, insufficient correlation modeling and missing reliability constraints in the reliability analysis of the truss structure, can significantly reduce the complexity of the truss size optimization design, improve the structure failure probability evaluation precision, realize efficient and accurate system reliability evaluation and lightweight design, and has good computability and engineering applicability while ensuring the precision.

The embodiment of the invention discloses a vehicle hardware fault tree analysis method and device, a medium and a program product. The method comprises the steps of determining a safety target for fault analysis of a vehicle based on a vehicle safety integrity level defined in a vehicle safety standard; taking the event violating the safety target as a top event, tracing at least one causal hardware chain between the hardware modules forming the top event, and constructing a fault tree; wherein a top end module of the causal hardware chain is an execution module which directly causes a top event, a middle module is a control module, and a bottom end module is a sensing module; and performing layer-by-layer detection on each causal hardware chain in the fault tree along the direction from the top end to the bottom end, and determining a minimum cut set causing a top event. According to the scheme, the fault tree can be constructed based on the causal hardware chain, and layer-by-layer detection is performed, so that the hardware modules are detected more logically and orderly, and the minimum cut set which may cause the occurrence of the top event is searched more comprehensively.

The invention relates to the field of fault tree analysis, in particular to a fault tree cut set solving method based on artificial intelligence. The method comprises the following steps: coding a fault tree into a structure matrix in a similar adjacent matrix form, and expressing an event and a cut set in a binary matrix form; searching the structure matrix to find out an analyzable intermediate event; analyzing each currently analyzable intermediate event through logic gate operation; performing a simplification operation on the analyzed intermediate event cut set, and only reserving a minimum cut set; after all parsable events searched in the round are parsed, rows and columns corresponding to parsed events in the middle event area of the structure matrix are deleted, and the next round of search is continued; and repeating the operation until the top event is analyzed. According to the method, the cut set is solved by adopting the matrix Boolean algebraic method, the fault tree analysis cut set simplification agent model is built based on the neural network, and the cut set is simplified in the solving process, so that the calculation efficiency of the minimum cut set is improved.

The embodiment of the invention provides a defense resource determination method and device, equipment and a medium, which are used for solving the problems that the accuracy of manual defense resource allocation is low and the security of defense equipment cannot be effectively improved in the related technology. In the embodiment of the invention, the electronic equipment maps the topological connection relation of the network into the undirected graph and takes the traffic of each link as the weight of the edge, so that the actual traffic distribution condition of the network can be intuitively and accurately reflected, and the minimum cut set is determined on the undirected graph by utilizing the minimum cut algorithm, so that the traffic distribution accuracy is improved. The electronic equipment can quickly identify the key link in the network. Based on the traffic of each link, the preset link parameter and the target function, the electronic device can scientifically predict the target newly-added defense resource quantity required by each link. The prediction method not only considers the current network flow condition, but also combines the characteristics and defense requirements of the link, thereby ensuring the rationality and effectiveness of resource allocation, and improving the security of defense equipment.

Embodiments of the present application provide a defense resource determination method and device, equipment and medium, which are used to solve the problem of low accuracy of manual allocation of defense resources and inability to effectively improve the security of defense equipment in related technologies. In the embodiments of the present application, the electronic device maps the topological connection relationship of the network into an undirected graph, and takes the traffic of each link as the weight of the edge, which can intuitively and accurately reflect the actual traffic distribution of the network. The electronic device can quickly identify the key link in the network by using the minimum cut algorithm to determine the minimum cut set on the undirected graph. Based on the traffic of each link, the preset link parameter and the target function, the electronic device can scientifically predict the target amount of newly added defense resources required by each link. This prediction method not only considers the current network traffic condition, but also combines the characteristics and defense requirements of the link, thereby ensuring the rationality and effectiveness of resource allocation and improving the security of defense equipment.

The application discloses a motor vehicle exhaust remote sensing station site selection method based on a minimum cut, and comprises the following steps: S1, converting a city traffic network into a directed graph, and searching the directed graph by a depth-first search algorithm to obtain a city road section set, wherein the city road section set comprises a directed loop road section, a directed non-loop road section and a boundary point distribution road section; S2, performing a point distribution operation on the directed loop road section; converting a motor vehicle exhaust remote sensing station site selection problem into a minimum cut problem, finding a minimum cut set based on a path segmentation algorithm of the minimum cut, optimizing by a local search algorithm, obtaining a result and outputting. The method only needs to use traffic network topological structure and traffic flow direction information, converts the motor vehicle remote sensing station site selection problem into a minimum cut problem, thereby obtaining a minimum road section set covering the traffic network, and enables all on-road motor vehicles in the city traffic network to be monitored at least once.

The invention discloses a parallelized cut set simplification method based on prime number compression coding, and relates to the field of fault tree analysis. In order to solve the problems that in the prior art, the cut set simplification calculation process is serialized, the parallel efficiency is low, a data structure is not matched with a GPU parallel architecture, and the minimum cut set of a large-scale complex system is difficult to efficiently solve, the method is provided to express the cut set in a fault tree model in a matrix mode in a binary vector form; grouping is carried out according to the number of events under the GPU parallel architecture; unique prime number codes are distributed for the basic events, the cut sets are converted into compressed integers, and division judgment of the inclusion relation of the cut sets is achieved; through matrix broadcasting and Boolean judgment matrix, large-scale parallelization subset identification is realized, and an effective cut set is dynamically updated in combination with an instant rejection mechanism, so that the calculation efficiency is remarkably improved, and the memory occupation is reduced. The method is suitable for system reliability evaluation and safety analysis work.

The invention discloses a matrix parallel computing method for solving a minimum cut set of a fault tree, and relates to the field of fault tree analysis. The matrix parallel computing method for solving the minimum cut set of the fault tree aims at solving the problems that in the prior art, the computing process is serialized, the parallel efficiency is low, a data structure is not matched with a GPU architecture, and the minimum cut set of the large-scale fault tree cannot be efficiently solved. According to the method, structured coding is carried out on a fault tree, and gate events and input relations are constructed into a structure description matrix; vectorizing the basic event and representing a candidate cut set of the intermediate event in a binary matrix form; under a GPU parallel architecture, row merging is performed on an OR gate, broadcast type combination is performed on an AND gate, parallel cut set simplification is performed, and meanwhile, a blocking and cache multiplexing strategy is introduced to improve the calculation throughput rate and reduce memory occupation. And carrying out loop iteration until the top event matrix is stable, and outputting a minimum cut set. The method is suitable for safety and reliability analysis of a large complex engineering system.

The invention discloses a quantification method of a minimum cut set of an earthquake PSA and a related device thereof, and the method comprises the steps: obtaining first data which comprises an earthquake risk curve, equipment object vulnerability information and equipment object random failure information; obtaining a target minimum cut set; a target binary decision diagram is generated based on the first data and the target minimum cut set, the data structure of the target binary decision diagram comprises a node probability, and the node probability is determined based on the first data; and based on the target binary decision diagram, determining a non-intersection cut set probability by using a recursive algorithm fused with a multiplication function, the multiplication function being related to a node probability. According to the scheme, the data structure of the binary decision diagram is optimized, and meanwhile, the cumulative multiplication algorithm is fused in the recursion process, so that the path probability can be obtained in real time in the recursion process, processing and calculation are not needed after a non-intersecting cut set is obtained, and the efficiency of the binary decision diagram algorithm and the accuracy of minimum cut set quantification are effectively improved.

The application discloses a fault positioning method, electronic equipment, a storage medium and a program product, and relates to the technical field of system operation and maintenance, and can determine the similarity of a target fault and a historical fault according to characteristic attributes of multiple dimensions, can avoid one-sidedness of similarity calculation based on a single dimension, can improve the accuracy of the similarity, and can further improve the accuracy of fault screening based on the similarity. Moreover, the fault positioning method constructs a fault tree corresponding to the historical fault, extracts a minimum cut set, and determines a candidate root cause of the target fault according to the minimum cut set, so that redundant checking can be avoided, and the fault positioning efficiency can be improved.

The invention provides a constellation system elasticity evaluation method comprising the following steps: S1, normalizing the coverage area of a constellation system in each working mode, and fitting to obtain a performance function; s2, determining the transfer type of the working mode, and determining the performance change condition of the constellation system before and after disturbance; s3, analyzing a minimum cut set Yk according to a preset fault tree; s4, calculating the elasticity of each minimum cut set Yk after an abnormal event occurs at the moment ti and a specific mode is implemented; s5, averagely taking m discrete time points from one work period of the constellation system, setting perturbation and the like to possibly occur at the m time points, and calculating the average projectile degree and the projectile degree satisfaction rate after a specific mode is implemented in the minimum cut set Yk; s6, calculating the average projectile degree and the projectile degree satisfaction rate of the minimum cut set Yk; s7, calculating the average elastic degree and the elastic degree satisfaction rate of the working mode transfer; and S8, calculating the average elastic degree and the elastic degree satisfaction rate of the constellation system.

The invention provides a cast-in-situ bored pile construction risk analysis method and system, and belongs to the technical field of risk identification, and the method comprises the following steps: determining a top event of a cast-in-situ bored pile construction risk by using an accident tree analysis method, analyzing an intermediate event and a basic event of an accident according to the top event, and constructing an accident tree; solving a minimum cut set of top events in the accident tree by using a Boolean algebraic method; drawing a dual tree of the accident tree by using duality of the minimum path set and the minimum cut set, and solving the accident tree by using a Boolean algebraic method to obtain the minimum path set of the top event; judging the structural importance degree of the basic event to the top event according to the occurrence frequency of the basic event in the minimum cut set or the minimum path set; and judging the risk factor of the top event through the minimum cut set, the minimum path set or the structural importance degree. According to the invention, the occurrence of accidents can be comprehensively analyzed, and prevention measures with strong pertinence are provided for avoiding the occurrence of accidents.

The invention discloses a nuclear power accident information processing method and device, and the method comprises the steps: obtaining N pieces of nuclear power accident information corresponding to a target nuclear power accident final state, and each piece of nuclear power accident information comprises an originating event and a function event; obtaining fault tree information corresponding to each piece of nuclear power accident information according to each piece of nuclear power accident information; calling a plurality of working nodes to carry out parallel processing on the N pieces of fault tree information to obtain a minimum cut set corresponding to each piece of fault tree information; and combining the obtained minimum cut sets corresponding to the multiple pieces of fault tree information to obtain a global minimum cut set of the target nuclear power accident final state, in the steps, parallel independent calculation of the nuclear power accident information can be performed, the overall efficiency is improved, the local minimum cut sets obtained through distributed calculation are efficiently combined, the global minimum cut set is generated, and the overall efficiency of the target nuclear power accident final state is improved. The required set can be quickly obtained, and subsequent safety evaluation is facilitated.

The application provides a system vulnerability attack disposal method, program product, storage medium and equipment. In the method, based on asset topology data and component vulnerability data of a target system, an initial attack graph is constructed, the influence of a security policy of asset configuration is considered, the attack success probability corresponding to each edge in the initial attack graph is determined, the initial attack graph is optimized, an optimized attack graph is obtained, then, a minimum cut set is solved in the optimized attack graph, and a minimum edge set capable of cutting all paths from an initial node to a final node, that is, a minimum disposal set, is generated. In this way, by analyzing the minimum disposal set on the constructed attack graph, the vulnerability disposal efficiency is effectively improved, the security operation cost is reduced, meanwhile, the attack graph is optimized according to the security policy of asset configuration, the information expression of the attack graph is more accurate, and the accuracy of vulnerability disposal is improved.

The invention discloses a mechanical arm vacuum circuit fault diagnosis and self-healing system and method, relates to the technical field of industrial robot adsorption execution, and solves the problems that vacuum circuit abnormity lacks a traceable diagnosis mechanism, fault recognition is inaccurate and no self-healing strategy support exists. In combination with AND gate / OR gate trigger logic and a minimum cut set recursive expansion technology, the vacuum system has a complete and clear fault conduction path, so that system-level anomaly can be accurately mapped to a bottom layer physical root cause, the structured traceability which cannot be achieved by a traditional method is realized, deviation caused by artificial experience judgment is effectively reduced, and the reliability of the system is improved. And the fault positioning accuracy and diagnosis efficiency are improved.

The invention provides a design method and device of a magnetically controlled reactor step type alternating current and direct current pulse hybrid controller, and belongs to the technical field of magnetically controlled reactor step type alternating current and direct current pulse hybrid controllers. An inverter circuit and a filter circuit are configured to realize controllable AC voltage output, a nanometer biphase composite magnetic material magnetizing control mode is established, a DC magnetizing process is realized through IGBT switch tube control, a graph theory minimum cut set-based stepped AC demagnetization optimization strategy is constructed, and stable magnetizing and demagnetization switching is realized by solving an optimal switch sequence. A self-adaptive control strategy based on a magnetization state prediction model is implemented, a hierarchical attention network is adopted to predict a magnetization state change trend in a magnetizing and demagnetizing switching process, and a dynamic parameter adjustment function is combined to optimize control parameters; the technical problem that accurate magnetization state control cannot be achieved in the direct current magnetization and alternating current demagnetization switching process of the magnetically controlled reactor is solved.