Vulnerability Assessment of Important Infrastructure of Power Grid to Natural Disasters Based on Deep Learning

As a typical complex network, the consequence of the accident is not simple equipment damage or regional power failure. Due to its cascading fault characteristics, some damages in the network may cause damage to equipment in other adjacent areas and lead to a wider range of power outages. Based on the theory of multi-attribute evaluation, complex network theory, and disaster science, this research qualitatively analyzes the composition and influencing factors of power grid vulnerability. It quantifies the multi-attribute composition of natural disaster vulnerability and then puts forward the corresponding evaluation methods and models. On this basis, the vulnerability of the power grid is comprehensively evaluated. The experimental results show that the natural disaster vulnerability assessment model of the important infrastructure of the power grid constructed in this paper considers the risk, reliability, and importance of the power grid. The identification accuracy of up to 80% shows that the evaluation path of power grid reconfiguration optimization and power grid optimal configuration given by the model can maximize the overall optimization of power grid operation.