Real-time Control Framework:


Grid Agent

The Grid Agent (GA) or power flow manager is a unique solution of GridSteer.
It takes care of power flows in the grid with many objectives.

Hierarchical Structure

The structure of a Commelec based control system is as follows: 

  • The Root Grid Agent (Root GA) is on top of the Commelec hierarchy. It is the leader of the overall grid controlled via COMMELEC.

  • The Internal Grid Agent (Internal GA) is the leader of a sub-grid. The resources in the sub-grid are followers of the Internal GA. The Internal GA is either a follower of the Root GA or of another Internal GA.

  • The Resource Agent (RA) is the end of a branch in the Commelec hierarchy. It interfaces the control framework with the physical resources of the grid that is being controlled (e.g., charging stations, boilers, photovoltaic plants, battery energy storage systems, etc).


Resource Agent (RA)

The Resource Agent is specifically developed for the resource it is interfaced with. However, different RAs have similar responsibilities:

  1. Communicate with the resource using its specific communication protocol:​​

    • ​To receive information from the resource (e.g., measurements, state transitions, etc)​​

    • To send the power setpoint computed by the GA to the resource so that it implements the provided setpoint

  2. Convert the internal state of the resource to a generic advertisement of the real-time capabilities of the resource and send it to a grid agent. The structure of the advertisement is resource independent. It contains:

    • What the resource can do​ 

    • A mathematical description of the uncertainty on the implementation of a given setpoint

    • What it would like to do

The two points mentioned above enable the abstraction of the resources in the grid. In other words, the Grid Agent does not know the type of the resources that it is controlling.


Grid Agent (GA)

The GA takes care of power flows in the grid with customizable objectives, energy resources and grid topology. It receives advertisements from all its followers (either resource agents or grid agents) and solves an optimization problem to dispatch the power flows in the grid under control.

Root Grid Agent

The root grid agent is the central element of our control framework. Its objectives are:

  1. Overall grid consumption follows a given aggregated pre-computed optimal power profile

  2. Maximize the fulfillment of the wishes of the resources

  3. Enforce safe operation of the grid

  4. Control the reactive power flow at the point of coupling with the distribution grid.

Internal Grid Agent

The internal grid agent is the aggregator of our control framework. It can be used to aggregate the advertisements of a group of resources and send the aggregated advertisement to the root grid agent. This principle allows for high scalability since it implies that entire sub-grids can be hidden behind a single internal grid agent.