Dynamic Bayesian network-based seismic resilience evaluation for ±800kV UHV converter stations

Ultra-high voltage (UHV) converter stations are essential to the seismic resistance and recovery capacity of areas without power after an earthquake. This paper uses seismic resilience as the focal point of dynamic Bayesian network (DBN), building a probabilistic seismic resilience assessment framework. First, the processes of damage and maintenance were mathematically modeled based on the theory of stochastic process. The effect of main and aftershock sequences on resilience was considered in the framework. A maintainability function was proposed to derive the spares’ consumption time and Markov transition matrixes. Second, a system functional model was established based on the topology and operation modes of a UHV converter station, which reflects the complex constraint dependencies among subsystems. Then, the time-varying analytical solutions for system functionality were derived based on Bayes' theorem. Finally, a typical ±800 kV UHV converter station was employed as a case study to demonstrate the applicability of the proposed framework. The effect of aftershocks and the effect of spare constraints on the system recovery process were studied. The proposed framework provides a method for evaluating the seismic resilience of converter stations from a probabilistic analytical perspective.