AMIVP - A Machine Intelligence based Storage Allocator for Virtual Power Plant Applications

B. Abegaz
Loyola University of Chicago,
United States

Keywords: power grids, energy storage systems (ESS), directional induction, virtual power plant (VPP), performance


A machine intelligence based storage allocator (AMIVP) is proposed to solve the problem of real-time integration and intelligent control of energy storage systems. Planning to utilize the available energy storage capacity for balancing and compensation of variations and perturbations from distributed sources, AMIVP computes the feasible storage allocation in a virtual power plant setting using temporal, forward and reverse directional inferences, taking the frequency of dispatch, state of charge, depth of discharge, over/under charge and leakage conditions of the storage system into consideration. The virtual power plant having sets of 200 V, 500 Ah Lithium-ion energy storage systems was modeled using Matlab-Simulink whereas the proposed directional induction method was developed in an IBM ILOG studio. The results of the implementation show that AMIVP could provide upto 49% improvements in the capacity, availability and performance of energy storage systems, thus enabling them to be used for virtual power plant applications. Suggestions to use the proposed allocator for voltage regulation as a means of paving the way towards a more sustainable and modern electric power grid are presented.