Planejamento de Baterias em Redes de Distribuição Inteligentes sob Incerteza

Abstract

Currently, power distribution systems are undergoing a transformation towards smart grids (SGs). One of the expected characteristics of the SGs is that they incorporate distributed energy resources based on renewable sources, especially solar. The lack of simultaneity between energy production of these sources and demand may result in operational issues, from which the risk of overvoltage stands out. Due to that, it is expected that energy storage systems (ESSs) play a key role in SG, since they allow to match the imbalances between demand and energy production, thus mitigating voltage violation issues. However, these devices are expensive, and therefore, the investments in ESS must be planned to obtain the best cost-benefit relation. In this context, a battery ESS (BESS) planning methodology for distribution SGs is proposed in this work, aiming to control the risk of voltage conformity indices violation subject to uncertainty regarding load, photovoltaic generation e BESS unavailabilities due to failures of their components or in the telecommunication network (TCN) used for their control. The methodology is based in a sequence of studies that result BESSs’ number, sites and sizes and the configuration of the TCN, while also allow the planning engineer to test improvements in the solution by using on load tap changer devices or or by increasing TCN’s availability. Four subproblems are used with this purpose: an optimal BESS allocation, a method for obtaining equivalent failure rates for the battery banks, the TCN optimal design and the SG’s voltage conformity probabilistic assessment used to test solution options face to uncertainty. The first and third problems are solved by genetic algorithms, while the second and the last are based on sequential Monte Carlo simulation (MCS) techniques. Embedded in the allocation model are an optimization model used to assess the typical BESS schedules and a useful lifetime prediction model based on that scheduling. The useful lifetime prediction is also used to fit time-to-failure distributions which ate the input of the MCS that generate BESSs’ equivalent failure rates which are coherent with the expected operational conditions. On the other hand, the MCS used to assess the voltage conformity uses not only the same schedule model used by the allocation but also two optimizations models which represent real time controls for BESS rescheduling and reactive power adjustments used for corrective voltage control. Results assessed in 33 node test system shown that the proposed methodology is capable of producing planning solutions that reduce the voltage conformity indices violation risk to tolerable levels.