Estimating sensitivity of interrupted energy and outage costs for customers in government, institutions and office buildings due to distribution grid failures using multilevel Monte Carlo technique

Customers in the government, institutions (GI) and office buildings (OB) sectors face substantial financial losses when their operations are interrupted by power distribution grid failures. Several factors affect the amount of interrupted energy, also known as Expected Energy Not Supplied (EENS) as well as the Expected Customer Outage Cost (ECOST). These factors include line/cable failure rates, restoration time, customer peak load, load and cost models, and the time of failure occurrence. Estimation results may also vary based on whether the customer distribution feeder is an overhead line or an underground cable. Therefore, this article focuses on a sensitivity analysis to determine the impact of above-mentioned factors on the EENS and ECOST for GI and OB customer sectors resulting from stochastic failures in power distribution networks. The GI and OB customers connected to the benchmark networks provided by Roy Billinton are taken into consideration. Sensitivities are estimated using an efficient and robust stochastic multilevel Monte Carlo simulation technique. A comparison of estimated outcomes for overhead and underground distribution systems is carried out. The findings show that ignoring variability in factors can lead to inaccurate planning and operational decisions related to EENS and ECOST for GI and OB customers.