Substation State Space Modelling and Reliability Assessment Considering Relay Protection Misoperation and Refusal

ABSTRACT This paper proposes a substation state space model considering relay protection misoperation and refusal, as well as an improved reliability assessment method in the stratified sampling stage. The aim is to evaluate the impact of substation load loss and the reliability of adjacent power grid lines during the fault isolation stage after the substation protection action. Firstly, based on the analysis of substation failure consequences and recovery strategies, it divides the time sequence of substation state space transitions, constructs the Markov state transition space of the substation affected by relay protection failures. Secondly, it puts forward a multi‐objective Monte Carlo stratified sampling method considering optimal allocation. By minimising the variance increment caused by the deviation between the actual allocation and the optimal allocation of the number of sampled states in each layer, including the main grid layer, the substation layer and the component layer, as the optimisation objective, it improves the results of traditional Monte Carlo simulations. Finally, this paper constructs a reliability assessment framework that combines the Markov model and Monte Carlo stratified sampling. It is verified through simulations in the modified IEEE‐RTS79 test system. The results show that after considering the relay protection intervention, the reliability indicators have increased to varying degrees. For the Monte Carlo stratified sampling method considering optimal allocation, taking the indicator EPNS (expected power not supplied) as an example, after 50,000 sampling states, the coefficient of variance has decreased by 95.2796% compared to the initial value. This indicates that, compared to traditional methods, this method can more accurately reflect the impact of protection misoperation and refusal on the reliability of substations and adjacent power grid lines. Moreover, while broadening the assessment boundary, it also provides higher computational accuracy and efficiency.