Dissertation defense (October 22, 2020): Gustavo Pacheco Epifanio

Abstract: This research has the objective of investigating the Switch Allocation Problems (SAP), which encompass a series of decision-making faced by power distribution utilities concerning the number, type, and location of switches in a network in order to reduce operational costs while maintaining acceptable levels of reliability. The Distributed Generation (DG) refers to the existence of an electrical generator positioned near the final user, allowing to make use of its energy generation. With this configuration, it’s expected to bring loss reduction in the electric network and also help to reduce the impact of system failure. The rationale is that DG may have a positive impact on reliability, given its potential to provide alternative power supply paths after a contingency through islanding operation. This research has the goal to investigate the SAP that encompasses a series of decision-making processes faced by the electrical distribution company to reduce operational costs and maintaining a predefined level of reliability on the distribution system. The solution is based on genetic and memetic algorithms. An optimization method is proposed to choose suitable values set of hyperparameters.  Computational experiments are conducted to evaluate the methodology on real-life networks. The results show the effectiveness of the approach to solve the PAC and the benefits in reliability that can be obtained from the use of DG.