We are developing a framework that enables the experimentation with grids in a realistic way. The decisions of the system are processed by a migrogrid tool called gridlabd. This tool performs static analysis of a microgrid to determine stable configurations and overloading situations. However, this tool is not designed to test facilities where decisions are made in run-time. So, we have devised a framework that interfaces with gridlabd to provide this functionality.

The proof of concept is being used in the MIRED-CON project, where a microgrid is enhanced using devices developed by ZiV company, tested by the University of Zaragoza, and deployed in the CEDER-CIEMAT facilities. UCM develops the intelligence part that will determine in each instant which powergrid setup is more efficient according to several, sometimes opposite, goals.

The current status of our development is showcased in this video. The video assumes there is energy consumption curve which needs to be satisfied either by power borrowed from a substation (this costs money since you are buying electricity) or by internal micro-generation. You count on several renewal energy sources, batteries, and others, some of which can be controlled. The system shows the effect of the orders generated by our intelligence and how these orders are keeping powergrid losses low and how the different generators are activated/deactivated to meet the demand. It should be said that, in some cases, it is not convenient to have an excess of energy, at least, according to spanish regulations. Hence, something detecting and avoiding this kind of situations, becomes necessary.

Download directly