I-Corps: Approximate Dynamic Programming and Artificial Neural Network Control for Microgrids

University Of Alabama Tuscaloosa, Tuscaloosa AL

Investigators

Abstract

The broader impact/commercial potential of this I-Corps project is to act as a catalyst in the growth of distributed generation and microgrid industries. This artificial intelligence based control system will potentially provide an electrical network that is reliable by reducing outages and restoration costs with incredibly fast bidirectional power flow, secured with real time diagnostics, self-healing and adaptive capabilities, and more economical by reducing equipment failures and minimizing power losses. The product potentially three broad markets, including utilities, distributed generation and consumer. The solution will enhance energy generation from renewables, improve microgrid efficiency, reliability, stability and power quality, and add intelligent control to conventional power systems. Inverter capabilities are presently a significant challenge for integrating distributed generation sources. The proposed innovation would potentially provide an appropriate solution to address this challenge. This I-Corps project develops a neural network control technology for microgrid control and management. Microgrids are one path for integrating renewable and distributed generation sources into the grid and can generally support a future smart electricity grid. A key challenge in microgrid adoption is adequate control of power inverters. Problems include high oscillations when connecting or disconnecting an energy source, fluctuating voltage and frequency, malfunctions and reliability, competing control between inverters, and high harmonic distortions. The proposed innovation uses adaptive dynamic programming and artificial neural networks to implement microgrid control. It integrates into one controller the advantages of conventional control methods, including optimal control, proportional integral control, predictive control, and sliding mode control. The proposed innovation has the potential to overcome the limitations of the conventional control technologies and better meet customer demands and requirements.