Multi-source data fusion method and system for dynamic system scenario behavior deduction and reliability prediction analysis

A multi-source data fusion system for dynamic system scenario behavior deduction and reliability prediction analysis, comprising:

• a particle swarm distribution acquiring module, configured to:
• obtain, based on a system state spatio-temporal coupling model, a particle swarm distribution of an assimilated initial system state by Gaussian sampling according to system state initial value input or assumption, import the particle swarm distribution to a system simulation model to simulate a trajectory change of a particle swarm, obtain a posteriori distribution of a system state particle swarm at the current moment by data assimilation in combination with a real-time monitoring input of a system state, and use the posteriori distribution as a particle swarm distribution of an assimilated initial system state at the next moment;
• a system state space model reconstruction module, configured to:
• determine, based on a range of the posteriori distribution of the system state particle swarm at the current moment obtained during data assimilation, an upper boundary value and a lower boundary value of a system state originating cell, use the upper boundary value and the lower boundary value as a new cell scale for discrete partition of a system state space, and complete reconstruction of a system state space model by digital coding;
• a Markov/CCMT system state transition probability mapping matrix model generation module, configured to:
• perform uniform sampling on the system state originating cell to obtain a distribution of a system dynamic behavior deduction initial state particle swarm, import the system dynamic behavior deduction initial state particle swarm into the system simulation model to simulate a trajectory change of the extracted particle swarm, obtain a conditional transition probability matrix of a system state under a current system configuration based on statistical estimation of a number of particle swarms falling into different cells, and obtain a Markov/CCMT system state transition probability mapping matrix model by combining the conditional transition probability matrix with a system physical component state transition probability matrix;
• a system state scenario behavior dynamic deduction algorithm and reliability prediction analysis module, configured to:
• perform, based on the constructed Markov/CCMT system state transition probability mapping matrix model, system state advance prediction analysis through applying a system state scenario behavior dynamic deduction algorithm, to further merge the same system states, push an exclusive system state obtained after merging and reduction into a search sequence structure, and repeatedly perform processes of construction, updating and iterative search of the Markov/CCMT system state transition probability mapping matrix model until reaching a specified search depth;
• and
• a system analysis result displaying module, configured to rank, based on the occurrence probabilities of the merged system states, predicted system state scenario sequence dynamic evolution and development, and display graphically through an interface to guide safe operation of the system.