Efficient Estimation of Interrupted Energy with Time-Varying Load Models for Distribution Systems Planning Studies

In electric power distribution systems planning studies, significant computational challenge is introduced when estimating customer interrupted energy which is generally referred as Expected Energy Not Supplied (EENS) index. In this paper, a computationally efficient technique for calculating EENS by considering time-varying load models (TVLM) is presented. The proposed technique is based on an advanced Monte Carlo (MC) method called the Multilevel Monte Carlo (MLMC), coupled with the Euler-Maruyama discretisation scheme for numerical solution of stochastic differential equation (SDE). In the traditional planning practice, an average constant load for a specific customer type is utilized in EENS estimation by ignoring different customer sectors TVLM. The proposed improvement of the practice is to consider typical daily load curves of different sectors to achieve better accuracy in the estimation of EENS. The proposed method is applied to the benchmark Roy Billinton Test System (RBTS) consisting of networks with five load buses and seven different customer sectors. The results show that the distribution systems with different load models produce dissimilar results. Also, the proposed method produces almost the same results as MC method and it is several times faster than original MC method.