Active Power Management in an AC Microgrid Using Artificial Neural Network and DSTATCOM

The evolution of non-renewable energy sources in modern era leads to the perception of Micro Grid (MG). The enormous energy demand of the exponentially increasing population requires alternative energy resources. The distributed resources are eco-friendly and emission less or nonhazardous in nature hence gaining more attention for present as well as future energy mandate. In this aspect the resources like solar, wind, geothermal and fuel cell energy are of prime importance because of their easy accessibility. The distributed and small scale energy production of MGs makes it very difficult for them to be tied and synchronized to utility grid. Further even after the coupling of MG with traditional grid system proper monitoring and maintenance of the power quality is indispensable. In this study the active power (P), reactive power (Q) and the Power Factor (PF) of a grid tied MG system is monitored and maintained with the help of Artificial Neural Network (ANN) trained Distributed Static Compensator (DSTATCOM). The superiority of the ANN is compared with that of traditional fuzzy logic controller (FLC). A brief analysis is executed by the simulation data of the proposed system, for multiple operating circumstances using Matlab/Simulink architecture. The results obtained confirms that the ANN controller is more efficient than FLC for improving the system characteristics in terms of better efficiency, stability, and dynamic response.