Risk Prediction of Power System for Typical Natural Disasters Based on FRAM and Dynamic Bayesian Network

The application of Bayesian networks in quantitative risk assessment is proved to be effective. However, interconnected cause network and visualization challenges grid manager when addressing multi-factor coupling accidents. In this paper, we combine the FRAM model and dynamic Bayesian network to forecast the risk evolution of power grid in the case of typical natural disasters. The FRAM model can effectively identify couplings between modules and provide the causal relationship for the topology of dynamic Bayesian network, which can quantitatively measure the effect of multi-dimensional factor coupling. The findings demonstrate that the smart grid provides a foundation for the successful application of the model to predict dynamic tripping, off-grid new energy, DC system faults, and other associated grid incidents.