Statistical characterization of reliability indices in medium voltage networks using a Monte Carlo-based method

Abstract

Distribution companies are compelled to provide high-quality services to their customers in compliance with existing regulations. Key performance indicators such as system average interruption duration index and system average interruption frequency index are commonly employed to assess the quality of service provided by these utilities. In certain countries, distribution system operators may face penalties or receive bonuses based on their performance within specified thresholds in their reliability indices. This paper introduces a novel, integrated algorithm, based on Monte Carlo simulations, for conducting a risk assessment in real MV networks using local data from the distribution system operator. The proposed technique is tested in a real MV network, and the obtained results are analyzed both at feeder level and for the entire network. The feeder-level results might be used to identify the sections of the network with the worst quality of power supply, based on the estimated KPIs. The overall results of the study offer valuable insights into the likelihood of specific KPI values using probability density function. The Weibull distribution and log-logistic distribution are found to be those with the best-fitting probability density functions for the two indices evaluated in this work. Furthermore, a penalty-based rate approach is presented as an application of the proposed method to assess the economic risk of the MV network. The findings reveal that the test system is expected to incur penalties amounting to 7.55 million dollars per year