Impact of correlated infeeds on risk-based power system security assessment

Abstract

In this paper we investigate how the detailed modelling of uncertainty impacts the assessment of the power system security. We propose a formulation for the inclusion of uncertainty in power system operational security assessment. The model proposed consists of a Monte-Carlo (MC) framework which allows to capture the complex stochasticity of the system inputs using the copula theory for the sampling of the system infeeds, in combination with AC power flow computations for a detailed assessment of the system operation. Taking into account the infeed forecast uncertainty (loads and renewable generation), and the set of possible contingencies, the MC formulation allows the detailed evaluation of the network operation by computing the system risk in terms of Lost Load. Moreover, the model computes the probability of cascading events, pointing out which remedial actions are to be expected and the associated amount of Lost Load. The proposed model is applied to the IEEE 118 bus test system. To highlight how uncertainty modelling impacts the results, we investigate how correlated infeeds affect the severity of estimated system stress. Results show that applying simplifications such as assumptions of independence or approaches that rely on DC approximations could lead to an underestimation of the system risk.