Senior Design Team sdmay21-24 • AI-VVO Cloud-based Machine/Deep Learning for Volt-VAR Control and Optimization

Project Statement

Distribution and regulation of energy is an essential issue in today's society. Historically, devices such as in-load tap changers, shunt capacitors, and in-line voltage regulators have been used to manage the voltage and reactive power of the target distribution grid. However, in general, these devices are slow and operate in discrete steps. Volt-VAR control (VVC) determines the strategies that these devices use. As distribution grids grow, they have to transmit large amounts of power over vast distances. In these situations reducing voltage fluctuations becomes key to reducing energy loss. Since current VVC devices are slow, they consume large amounts of power when voltage changes become more frequent. Recently, however, Smart inverters have been emerging as popular devices to assist with VVC. Researchers have begun exploring how machine learning could be used with these smart inverters for Volt-VAR optimization (VVO) to minimize energy loss, voltage deviation, and peak load.

This project aims to design and implement cloud-based machine learning or deep learning algorithms for VVC and VVO for distributed energy resources integrated distribution grid to increase voltage stability and reduce energy loss.