157 results about "Parallel simulation" patented technology

The invention relates to the technical field of artificial intelligence and automatic driving, and discloses an automatic driving digital twin agent system based on reinforcement learning training, which comprises an agent library construction module, a simulation and configuration module, a distributed simulation execution module, a scheduling and learning module, a strategy training module and a simulation migration module. The method comprises the following steps: constructing a modular digital twin agent library, combining to generate an urban traffic simulation scene, executing large-scale parallel simulation on a distributed computing cluster, implementing dynamic task scheduling and course learning according to an agent learning progress, and training a reinforcement learning strategy and a value network based on simulation data. And verifying the strategy model and implementing migration deployment from simulation to reality. According to the method, the training efficiency can be improved, the diversity of simulation environments is enriched, and the sim-to-real migration difficulty is reduced, so that the robustness of the strategy model is improved.

The invention discloses a digital twinning-based intelligent simulation control system for a power system, and the system comprises a data processing module which is used for collecting and preprocessing the operation data of the power system; the storage mapping module is used for mapping the data to the virtual architecture and generating a twin snapshot sequence; the factorization module is used for executing graph embedding and node decomposition to generate embedding features; the twinning mapping module is used for constructing a twinning mapping network and realizing bidirectional fusion; the simulation modeling module is used for executing health evolution and power flow risk modeling on the GPU to generate a knowledge graph; the strategy solving module is used for carrying out constraint calculation on the knowledge graph and screening an optimal strategy; and the deduction regulation and control module is used for generating a causal path based on graph backtracking and completing closed-loop control. Real-time mapping, intelligent deduction and self-adaptive regulation and control of the power system are realized through digital twin modeling and GPU parallel simulation, and the scheduling precision, the response speed and the operation stability are remarkably improved.

The invention discloses a real machine deployment method and device for a quasi-direct-drive robot for migration from simulation to reality, and the method comprises the steps: S1, driving each joint motor through a PD controller on a real machine of the robot to follow a plurality of groups of reference tracks, and synchronously collecting the real motion track data of each joint; s2, in a parallel simulation environment, performing initialization based on a random frame in a real motion track, randomly endowing a group of motor kinetic parameters, then driving a simulation robot through a PD controller by using the same group of reference tracks, and generating corresponding simulation motion track data; s3, based on a trajectory error between the two types of trajectory data, iteratively updating motor kinetic parameters by using an optimization algorithm until the error converges, thereby obtaining a calibrated simulator; s4, training a control strategy network of the robot through reinforcement learning in the calibrated simulator; and S5, directly deploying the control strategy network obtained by training to a real machine for performance evaluation and security verification.

The invention discloses a chain type hybrid deep reinforcement learning quadruped robot control system. The system comprises a sensing module, a data processing module, a chain type hybrid reinforcement learning model module, a feedback module and the like. The invention further discloses a control method of the chain type hybrid deep reinforcement learning quadruped robot control system, and the method comprises the steps: S1, collecting bionic motion data, robot state data and environment data, and training based on an attention mechanism space-time diagram neural network and antagonistic imitation learning to obtain a reference gait model containing motion prior; and S2, constructing a three-level parallel simulation environment, setting a progressive course target, training the progressive course target by adopting an improved PPO algorithm, and optimizing the gait stability, the movement efficiency and the environmental adaptability. By means of high-precision data acquisition with the sampling frequency larger than or equal to 240 Hz and antagonistic imitation learning, motion priori knowledge is introduced, the training period can be shortened by 50% or above, abnormal gaits are avoided, and the training process is controllable.

The invention belongs to the technical field of computers, particularly relates to an enterprise consultation service system based on big data, and aims to solve the problems of data islands, decision lagging and insufficient predictive ability in traditional enterprise consultation service. The system comprises a global data access module, a knowledge graph construction module, a dynamic simulation engine, an intelligent consultation generation module and an interactive feedback optimization module, a time sequence evolution knowledge network is constructed by fusing internal and external multi-source heterogeneous data, multi-scene parallel simulation and intelligent suggestion generation are realized, and closed-loop feedback learning is supported. According to the scheme, the accuracy, foresight and response efficiency of enterprise strategic decision can be improved, and large-scale concurrent service and real-time dynamic optimization are supported.

The invention discloses a real-time monitoring and early warning method and system for hoisting operation of a wind generating set, and relates to the technical field of wind generating set construction, the method comprises the following steps: establishing a digital twin model of a hoisting operation site, collecting state data in real time for dynamic coupling, and constructing a virtual monitoring network; performing short-term advanced parallel simulation in the network to obtain future state trajectory clusters under different potential interferences; extracting a key safety index time sequence from the trajectory cluster, and calculating a future instability risk probability; and when the risk probability exceeds an adaptive early warning threshold, generating multi-level early warning information and triggering active intervention. The technical problem that the instability risk is difficult to recognize and intervene in time due to the fact that the hoisting operation of the existing wind generating set lacks real-time prediction of the system state in a complex environment is solved, and the technical effects of real-time early warning and effective intervention in the hoisting process are achieved by constructing the digital twinborn monitoring network and carrying out risk propagation prediction.

The invention relates to the technical field of power system communication, and particularly discloses an optimization method for an operation and maintenance strategy of a power communication network, which performs space-time alignment and feature reconstruction on multi-source heterogeneous telemetry data of the power communication network by using a sliding window and a graph neural network, and constructs a state tensor reflecting network physical and logic panorama. Generating a candidate strategy set through a deep reinforcement learning action network, and mapping the candidate strategy set to a digital twin environment synchronized with the current network in real time for parallel simulation rehearsal; a dynamic scoring mechanism including network pressure sensing and marginal utility analysis is introduced, fusing is carried out on a high-risk strategy in combination with a logarithmic barrier function, and an optimal strategy giving consideration to performance and safety is selected from a candidate set. And finally, a final execution instruction is generated through verification of a power service hard rule filter, the executed real network is fed back and stored in an experience playback buffer area to update model parameters, and the problems that a traditional static strategy cannot dynamically adapt to the network pressure and the direct trial and error risk is high are solved.

The invention discloses a stake car wind resistance dynamic optimization method and system based on airflow simulation, and belongs to the field of automobile pneumatic optimization. A stake car three-dimensional structure model is constructed, and spatial distribution parameters and surface roughness parameters of a stake component are extracted; establishing a multi-working-condition external flow field simulation model, and obtaining a unit area wind resistance coefficient distribution diagram; constructing a wind resistance sensitivity matrix, and identifying a high-sensitivity component; performing disturbance modeling on the high-sensitivity component, generating a local deformation configuration sample and establishing a simulation model set; parallel simulation calculation is carried out, and wind resistance coefficient and aerodynamic stress data are extracted; based on a multi-objective optimization algorithm, determining an optimal disturbance configuration; outputting a structure adjustment suggestion for guiding dynamic design optimization; according to the method, collaborative optimization of wind resistance reduction and structure safety can be achieved, and the pneumatic performance and the energy-saving level of the whole stake car are improved.

The invention relates to the technical field of electromagnetic transient simulation, in particular to an electromagnetic transient simulation decoupling method, system and device based on a single-impedance branch and a medium, and the method comprises the steps: dividing a system to be simulated into two subsystems which are connected through the single-impedance branch, and carrying out the equivalent reconstruction to form a reconstructed branch model; establishing a discrete state space equation based on the reconstructed branch model; according to the main step length, the number of sub-steps meeting the stability condition is determined based on a small-step synthesis algorithm, and the main step length is equally divided into differential step lengths; in each main step length interval, performing iterative calculation on the state-space equation by using a differential step length to obtain a branch current and a capacitor voltage at the next moment; and injecting the calculated branch current serving as a known current source into an original network equation, solving in parallel to obtain the port voltage of the subsystem, and completing decoupling calculation and promoting simulation. According to the method, effective decoupling of the universal single-impedance branch can be realized, and the applicability, the numerical stability and the calculation efficiency of parallel simulation are improved.

The invention relates to the technical field of engineering project management, and discloses an engineering project progress and resource collaborative management system based on a cloud platform. According to the system, a progress main sequence is generated by collecting progress data, and resource information is recorded in real time to generate a resource circulation record set. The system extracts progress key nodes and corresponding resource consumption wave crest features and performs time alignment verification to generate preliminary cooperation deviation information; and the system inputs the information into a collaborative deduction model, and carries out parallel simulation on the subsequent progress and the resource demand to obtain a future collaborative deduction picture. The system analyzes on-site image data according to the image, obtains actual cooperative state parameters, and generates a management conclusion for adjusting subsequent plans and resource allocation by comparing the actual cooperative state parameters with deduced images. According to the invention, dynamic accurate diagnosis and look-ahead closed-loop regulation and control of the progress and resource coordination state are realized.

The invention discloses a smart city power marketing service simulation evaluation method and system based on digital twinning. The method comprises the steps of obtaining power marketing service data, performing space-time semantic analysis, extracting attribute dimensions, establishing a cross-domain attribute association graph, establishing a cross-domain attribute association graph, and constructing a digital twinborn model; the target marketing service process is decomposed into service units, the service units are decomposed, a local simulation domain is constructed for parallel simulation, and a combined simulation result is obtained through an inter-domain coupling operator; and comparing a simulation result with real data, identifying virtual and real deviations, performing reverse traceability decomposition on virtual and real deviation distribution, mapping the virtual and real deviation distribution back to each local simulation domain, determining a local contribution magnitude of each local simulation domain, and constructing a hierarchical evaluation index to obtain a service quality evaluation result. According to the invention, accurate evaluation of the power marketing service is realized, and the service quality supervision efficiency is improved.

The present application belongs to the field of photoelectric countermeasure simulation and digital twin technology, and relates to a laser jamming full-link digital parallel simulation system and method. The system comprises a scene initialization unit, a laser parameter configuration unit, a target characteristic definition unit, an atmospheric transmission calculation unit, a load imaging simulation unit, a jamming effect evaluation unit and a visualization terminal connected in sequence. The method realizes digital simulation of the whole process of laser jamming through the whole process coupling of scene configuration, atmospheric transmission calculation, load imaging simulation, jamming effect evaluation and visualization output. The present application solves the problems of lack of full-link integration, insufficient dynamic process simulation, weak semi-physical verification capability and poor scalability in the prior art by constructing a full-link integrated simulation architecture covering ground radiation, atmospheric transmission, load imaging and image quality evaluation, and can be widely applied to photoelectric countermeasure equipment effectiveness evaluation, countermeasure strategy optimization and equipment performance verification.

The embodiment of the invention provides a parallel simulation method, system and device of a multi-core processor and a storage medium, the parallel simulation method of the multi-core processor comprises the following steps: creating a context for each of a plurality of cores of the multi-core processor, the plurality of contexts corresponding to the plurality of cores being mutually independent; a system thread is created for each context of the plurality of contexts such that the plurality of cores are allocated in different contexts into different system threads that are automatically scheduled to different processor cores for execution. According to the parallel simulation method of the multi-core processor provided by the embodiment of the invention, the parallel simulation is realized based on a plurality of independent contexts, the problem of low simulation efficiency in a traditional simulation framework is effectively solved, and the execution speed of the simulation is remarkably improved through a multi-thread simulation technology.

This invention discloses a method and system for parallel training, evaluation, and control based on a digital twin using an equipment bus. It unifies the time-stamping and state-space mapping of multi-bus observation data, and solves for a unified state estimate based on quality indicators, fusion residuals, and cross-bus physical consistency constraints. This drives the state evolution of the digital twin, executing the Lyapunov descent admission condition under synchronization error and end-to-end delay constraints. Further, multi-scenario parallel simulations are conducted to obtain an upper bound on the risk probability and construct a training evaluation function. When the evaluation is below a threshold or the risk exceeds the limit, only the affected control components are subject to safe feasible region projection correction, and closed-loop control commands are output. This invention improves state estimation accuracy, synchronization stability, risk prediction capability, and training control safety.

The invention discloses a modular multilevel converter decoupling electromagnetic transient modeling method and a related device, and relates to the technical field of power electronics, and the method comprises the steps: determining equivalent parameters of a single-phase Bergeron transmission interface used for decoupling between a main network side and an MMC side based on an inductor, a resistor and a simulation step size which are connected in series with MMC bridge arms; based on the equivalent parameters, after an oscillation suppression branch is constructed on the MMC side and the parameters of the oscillation suppression branch are determined, the parameters of the remaining branches are determined; deriving electromagnetic transient formulas of the main network side and the MMC side, a circuit formula of the oscillation suppression branch and a circuit formula of the residual branch based on the parameters; and based on the electromagnetic transient formula, the circuit formula of the oscillation suppression branch and the circuit formula of the remaining branch, electromagnetic transient calculation of the main network side and the MMC side is decoupled, so that the main network side and the MMC side are subjected to independent parallel simulation calculation. According to the invention, the problems of high calculation complexity and low simulation reduction in the prior art are solved.

The invention discloses a period precise simulator design method for a combinable core particle network, and relates to the technical field of core simulation testing, and the method comprises the steps: designing a simulation frame of a combinable core particle network simulator; wherein the simulation framework comprises a two-stage core particle network configuration unit and a multi-thread parallel simulation framework; carrying out combinable topology modeling, modular routing mechanism modeling and heterogeneous router micro-architecture modeling on the combinable core particle network; performing protocol layer protocol conversion modeling, protocol layer flow control mechanism modeling, adaptation layer retransmission mechanism modeling and physical layer electrical behavior modeling on the core particle interconnection protocol interface; based on the above design, the period precision simulator for the combinable core particle network is constructed. By designing the simulation framework, more accurate actual core particle network behaviors can be obtained; a higher simulation speed is realized through a multi-thread parallel simulation framework, multi-thread parallel accelerated simulation under a large-scale network is supported, and the simulation test efficiency is improved.

The invention relates to the technical field of microelectronic assembly equipment control, and discloses a chip pin alignment calibration method based on machine vision, comprising the following steps: establishing a real-time motion error model representing mechanical positioning errors, and obtaining pin polar coordinates through partition extraction; a tangential margin is deducted from the theoretical region to construct a static tolerance window; performing morphological erosion and parallel simulation on the static window at a multi-stage speed by using an error boundary output by the model, and calculating a full-pin dynamic window geometric intersection; and according to the intersection non-empty state, the optimal speed and the target coordinate are decided, equipment mounting is driven, and thermal drift parameters are updated based on an observation residual closed loop. According to the method, the real-time motion error model containing thermal drift compensation is constructed, dynamic morphological erosion based on the mechanical positioning error boundary value is executed on the static window, the effectiveness of mounting coordinates in different thermal states is guaranteed, and dynamic decoupling of precision and speed is achieved.

A multistage fault-tolerant parallel multi-objective optimization method for a process simulation platform comprises the following steps: executing sandboxed isolation and off-peak disturbance starting; establishing a multi-level fault-tolerant and self-healing control mechanism, setting a task-level fault-tolerant mechanism, an instance-level fault-tolerant mechanism and a system-level fault-tolerant mechanism, and performing health examination, overtime monitoring and abnormal restarting on each parallel simulation instance; performing constraint perception and penalty fusion processing, performing linear and interval constraint screening at an optimization end, performing process constraint verification at a simulation end, and performing penalty correction when any constraint is not met; and normalization and diversity maintenance irrelevant to the algorithm: collaborative decision and result judgment are carried out, a comprehensive efficiency function and dominance degree and coordination degree combined evaluation model is constructed in a multi-dimensional normalization index space, and a representative solution with the most balanced performance and the optimal energy efficiency is determined.

This invention provides a method, apparatus, device, and storage medium for multi-branch parallel simulation and decision optimization based on a large language model. It constructs a dynamically evolving battlefield state space by receiving multi-source real-time situational data. Situational analysis identifies battlefield emergencies, which are then used as triggers to invoke the large language model to generate multiple differentiated decision branches, each containing macro-level objectives and micro-level strategies. For each decision branch, an independent parallel battlefield simulation instance is constructed and allocated computational and simulation resources, driving synchronous parallel simulations to obtain real-time simulation states. Multi-dimensional dynamic evaluation is performed to obtain a dynamic comprehensive score. Based on this, branches with scores continuously below a preset threshold are dynamically pruned, and their released computational and simulation resources are weighted and redistributed to surviving branches to improve their solution accuracy. Finally, high-scoring, high-quality branches are extracted, and the large language model is invoked to perform decision fusion on their key action sequences, outputting the optimal combat path and command sequence.

The application provides a floating slab track steel spring stiffness vehicle-mounted identification method based on digital twinning, relates to the technical field of rail transit, and comprises the following steps: performing small-area digital twinning calibration on vehicle dynamic response data, performing vehicle-mounted extension identification on the calibration result, and obtaining a line finite element model; performing collaborative hierarchical updating on a rail transit vibration model according to characteristic frequencies, updating track irregularity parameters in combination with axle box acceleration signals and Chinese spectrum, and obtaining a small-area digital twinning model; constructing a stiffness identification database according to the principle of floating slab steel spring stiffness degradation and damage; constructing a particle independent parallel simulation environment based on the stiffness identification database and vehicle-mounted acceleration signals, constructing an improved particle swarm optimization algorithm, and identifying the vehicle-mounted damage degree and position of the floating slab track steel spring stiffness based on the optimal parameter solution of the improved particle swarm optimization algorithm. The application solves the problem that the health state of the floating slab track steel spring cannot be grasped in time.

The invention relates to the field of engineering simulation, in particular to an iron roughneck oil cylinder hinge point driving optimization method and device, equipment, a medium and a product. According to the method, the basic simulation model capable of being regulated and controlled in a parameterized mode is established, the hinge point of the push-pull oil cylinder is set as a design variable, and the boom is subjected to flexible processing, so that the problem that vibration of the boom based on a rigid body cannot be simulated is solved, and driving function parameters can be dynamically adjusted along with the position of the hinge point; by initializing a parameter group and carrying out parallel simulation, the single optimization period is greatly shortened, and the driving parameters are prevented from being manually adjusted for multiple times; meanwhile, a dynamic graph is drawn in combination with simulation data and iterative data, and performance differences under different driving functions can be visually presented, so that the optimal solution of the oil cylinder force is accurately explored, and the defects that in the prior art, the optimization period is long, and the deviation between the result and the actual working condition is large are effectively overcome.

The application relates to a parallel multi-objective optimization method and system for engineering product CAE simulation, and belongs to the technical field of simulation optimization. The method solves the problems that the simulation-driven design cycle is long and it is difficult to obtain a high-quality multi-objective trade-off optimization scheme. The method comprises the following steps: based on multiple objective functions and constraint conditions, initial sample data is obtained through two-stage collaborative optimization and is put into a sample library; a joint surrogate model is constructed and trained, and a hierarchical error compensation mechanism is initialized; multiple rounds of iterative optimization are performed until a preset optimization termination condition is met; each round of iterative optimization comprises the following steps: based on the current joint surrogate model and the hierarchical error compensation mechanism, multiple candidate design points are generated in parallel, and then the multiple candidate design points are distributed to multiple computing nodes for parallel simulation calculation to obtain new sample data which is put into the sample library; the joint surrogate model and the hierarchical error compensation mechanism are updated; and a multi-objective optimization design scheme set of a product to be optimized is obtained from the final sample library. The optimization efficiency is improved.

The application relates to the field of simulation technology, and particularly discloses a heterogeneous interconnection multi-core distributed simulation experiment system and a cooperative control method. The system comprises a semi-physical simulation system and a virtual system. The semi-physical simulation system comprises a plurality of semi-physical simulation nodes, each of which can simulate a semi-physical simulation entity; a plurality of multi-core parallel simulation nodes, each of which can simulate a plurality of parallel semi-physical simulation entities. The virtual system comprises a digital simulation system, which can simulate a plurality of parallel digital simulation entities; an experiment environment model, which is used for generating simulation experiment environment modeling data; a virtual-real combined simulation experiment platform, which is used for deducing and rendering a working scene; and a dynamic deducing visual simulation engine, which is used for rendering environment information by using the simulation experiment environment modeling data. The system has the capability of multi-core simulation, can further improve the upper limit of semi-physical simulation model access to a semi-physical system, and enables the semi-physical system to perform cluster simulation.

The invention discloses a reservoir group risk adaptive scheduling method and system for coupling ensemble forecast uncertainty, and the method comprises the steps: fusing monitoring and meteorological ensemble forecast data in real time, and outputting an initialization field through assimilation processing; a complete scene set is generated through parallel hydrological-hydrodynamic simulation, and a dynamic risk probability field is calculated and clustered to obtain a weighted typical scene set; then, a rolling optimization model containing dynamic probability constraints is constructed, and an optimal instant scheduling instruction is solved and output; and updating data in a rolling manner according to a fixed period and repeating the process. The system adopts a modular design and comprises a data synchronization assimilation module, a parallel simulation and risk quantification module, a rolling optimization decision module, an instruction execution and feedback control module and a system support layer. According to the invention, the scheduling strategy adapts to the risk dynamic change in real time, the safety and benefit are balanced, the scheduling intelligence and standardization level are improved, and the engineering potential safety hazard is reduced.

The invention discloses a hybrid power system continuous dynamic-discrete strategy double-layer parallel simulation architecture which comprises a plurality of modules distributed in a master process and a slave process, the master process is used for simulating and solving a continuous subsystem model and controlling an overall simulation time sequence, and the slave process is used for simulating and solving a discrete subsystem model. The master process comprises a clock control module, a discrete event processing module, a continuous model grouping parallel module and a continuous state sending module, the slave process comprises a continuous state receiving module, a discrete model grouping parallel module and a discrete event sending module, and a shared memory module is set as an interaction channel of the master process and the slave process. According to the method, the simulation efficiency of the hybrid power system is effectively improved on the premise of ensuring the accuracy of a simulation time sequence through consecutive discrete subsystem parallelism and model grouping parallelism in the subsystems. The framework can be applied to hybrid power system simulation fields such as cascading failure analysis, power system stability control system checking and the like.

The invention relates to the field of traffic control, in particular to a regional traffic situation perception system based on fog computing, which comprises a perception acquisition module, an edge analysis module, a regional cognition module and a situation service module. The system acquires data through the sensing acquisition module; the edge analysis module constructs a digital twinborn model, and outputs a structured scene risk signal by using a multi-modal dynamic attention generative adversarial network and an intention reasoning model; the regional cognition module performs parallel simulation in the digital twin sand table and preferentially generates a collaborative strategy; and the system finally realizes closed-loop optimization of the internal model through a complete scene information packet formed by feedback. According to the invention, by constructing a cloud-edge collaborative digital twinning and multi-agent decision framework, intelligent optimization transformation of traffic management from passive response and local control to active pre-judgment and regional collaboration is realized.

The application relates to the technical field of air transportation management and intelligent decision-making, and specifically provides an airport cluster integrated scheduling method based on parallel simulation, aiming to solve the problems of poor real-time performance, inability to balance multiple targets and lack of decision-making closed loop in the prior art. The core of the method is as follows: a digital twin model of a physical airport cluster is constructed and continuous time sequence evolution is carried out; preset scheduling events are monitored in real time during the time sequence evolution; once an event is triggered, a target scheduling strategy is generated based on the current state through a parallel simulation and decision-making fusion process; finally, the strategy is mapped back to the digital twin model to replace the current logic to continue the time sequence evolution. Through the parallel system architecture of virtual and real interaction, online, real-time, self-adaptive optimization and decision-making support for the operation of the airport cluster are realized, and the scheduling efficiency, fairness and system resilience are significantly improved.

This invention discloses a method and system for dynamic twin closed-loop parallel simulation and extrapolation of safety warnings for autonomous driving. The method includes: acquiring multi-source data from real traffic scenarios and constructing a dynamic twin simulation environment that evolves synchronously with the real scenario; generating multiple future scenario evolution branches based on current scenario state and trajectory prediction information, and performing ultra-real-time parallel extrapolation to obtain scenario extrapolation results at multiple future time points; determining the consistency between the extrapolation results and real observation results, and dynamically pruning branches that do not meet the consistency conditions; automatically pruning branches with low threat levels based on threat level indicators; interacting with the onboard autonomous driving control system and the dynamic twin simulation environment to form a virtual-real closed-loop feedback; and outputting safety warnings, control suggestions, and safety decisions based on the pruned extrapolation results. This invention achieves ultra-real-time extrapolation, dynamic correction, and proactive warning of potential risks in complex traffic scenarios, improving the real-time performance, accuracy, and reliability of autonomous driving safety warnings.

This invention relates to the fields of automated robot technology and coal yard safety management technology in thermal power plants, and particularly to a coal yard debris cleaning device based on an automated robot. The device includes a sensing module that collects and fuses data to output a comprehensive data stream. A twin module dynamically constructs and updates a dynamic digital twin through coupled discrete element method and computational fluid dynamics simulation. A risk prediction module uses a graph neural network surrogate model and Monte Carlo simulation to generate a spontaneous combustion probability spatial distribution map as dynamic risk information. A decision-making module integrates dynamic risk information and robot state to generate a set of candidate joint action strategies. A control module verifies the strategies through parallel simulation and performs multi-objective arbitration based on dynamic risk information to select the optimal control law and decouple it into low-level control commands. A feedback module executes commands and collects actual data, feeding it back to the twin module to correct the dynamic digital twin and to the control module to optimize decision weights.

The invention relates to the technical field of real-time simulation, and particularly discloses a cascade H-bridge PET multi-device collaborative real-time simulation method and system based on an FPGA, and the method comprises the steps: positioning an electrical connection point between a DC side capacitor of each H-bridge unit and a DAB input interface in a PET system, splitting the system, and obtaining a three-phase cascade H-bridge part and each phase DAB DC level; the main simulator side simulates a three-phase cascade H-bridge part, adopts 3 * n groups of controlled current sources to be equivalent to a slave DAB direct current level, and measures the voltage vector at each controlled current source end in real time; and the three slave simulators adopt each phase cascade H-bridge sub-module of n groups of controlled voltage source equivalent hosts to synchronously measure the current vector of the controlled voltage source, and after processing, bidirectional data interaction between the master and slave simulators is realized through optical fibers. According to the method, the contradiction between a large-scale cascaded H-bridge system and high-frequency switch simulation is solved through heterogeneous step lengths, a parallel simulation architecture and a flexible signal receiving mechanism, and accurate real-time simulation of the cascaded H-bridge PET system is realized.