Mas Roig, Girona (Spain)

The pilot Microgrid is located in Mas Roig, Girona, Spain and comprises of a country house, a small farm and a number of renewable energy generators (PV panels, small wind turbines, batteries). The scope of this pilot site is to demonstrate that the local networks of renewable energy sources can operate and sustain themselves independently of what happens in the rest of the power grid. In case of grid imbalances (voltage or frequency deviations), Microgrids can support the grid operation by offering load management services based on a prioritized list of loads (lights, washing machines, heating/cooling devices etc). In the case of emergency grid conditions (islanding operation), Microgrids can operate autonomously and serve the energy needs by exploring exclusively the local renewable generation resources.

Image

Communication Infrastructure

The different elements into the infrastructure are communicated by means of ZigBee technology, with the inclusion of the so called ZigBee nodes. The use of the ZigBee, implies a great flexibility on the inclusion of new elements into the system.

Image

Each node is associated to elements and is adapted to its particular needs. The different node types are based on the different loads and generation characteristics, and can be summarized within the following groups:

  • Central Node: this node has an USB connection and is attached directly to the central computer.
  • RS485 communication: this node provides a communication from the central computer to the device where the ZigBee node is attached and which includes an RS485 port. Through this device, the central computer can read from that device by means of a ModBus protocol (Inverters and others).
Image
  • RS232 communication: this node provides a communication from the central computer to the device where the ZigBee node is attached and which includes an RS232 port. With this device, the central computer can read from that device by means of a ModBus protocol.
  • Control node – Static: this node provides the measurement on V and I on the load attached to the device. It has the capability to measure also to control the switching of a load up to 16 A. This node is always present in the system. Has to be registered and identified.
  • Control node –  Mobile: this node provides the measurement on V and I on the load attached to the device. It has the capability to measure also to control the switching of a load up to 16 A. Anyway, this node has the capability of appear and disappear into the system (as an example, an Iron, which can be plug in or plug out). When plug in, the device appear into the system automatically. In case of plug out, the node disappears. It doesn’t need to be registered before.

The selected topology for the Mas-Roig Microgrid is Star type. With this topology, the consumption and delays are the lower as possible, because every element is only active when transmitting and they don’t need to send the information packages to each other nodes. The framework technology used for the ZigBee infrastructure is Free-Scale Stack, and the time latency on reading of all nodes will be of 2 seconds, although we will test different sampling rates, due to the fact that the sampling rates has a strong relationship with node consumption, and it is a relevant point for this application.

Control Architecture

The architecture proposed is based on Multi-Agent Systems (MAS), in order to manage and coordinate efficiently the power system of the Mas Roig Micro-grid. The objective is combining the use of artificial intelligence algorithms (like CBR, reinforcement learning…etc) with the traditional computational methods to cope all the diverse problems faced in Micro-grids control.

The general description of the MAS system that allows us to control the Micro-grid power consumption is:

  • MGCC: (Micro-grid Central Controller Agent) is the main responsible for the optimization of the Micro-grid operation. MGCC controls the power exchange between the Microgrid and the upstream network as well as the provision of ancillary services to the system operators.
  • PV: a Production Unit Agent that is dedicated to represent the photovoltaic panel.
  • Wind: a Production Unit Agent that is dedicated to represent the small wind turbine.
  • Battery unit: a Production Unit Agent that is dedicated to represent the battery panel. This agent can sell or buy energy from the market depending on the state of charge of the batteries.
  • Power system: a Grid Agent that represents the grid. It should be mentioned that the grid is considered as a production unit or consumption unit with infinite capabilities.
  • Load unit: a Load Unit Agent that is specialized in representing the loads of the system. Each load Agent controls 1 or more ZigBee nodes.

Other agents in the architecture are the following ones:

  •  CA: Communication Agent. It provides the communication between the MAS system and the ZigBee net and distributes the information to the different LC.
  • UA: User Interface Agent: this agent provides the user interface and usability of the system for the Integral experiments. The number of different situations programming and situations registering will be tracked and managed by this agent. It has associated a Data Base for registering purposes.
  • DBA: Data Base Agent. Provides communication between the agents that needs access to data base and the database itself.
  • Data Base: The Data Base that appears into the architecture has a direct connection between Data Base and Internet. This is due to the fact that due to security reasons the only communication between the simulator and the MAS system is by interchange on registers data with a static IP address.

All these agents agents run under JADE platform in a unique PC under Windows and using Java language.

MAS Graphical User Interface (MAS-GUI)

A user-friendly graphical interface was developed which enables the monitoring of the Microgrid operation. The MGCC agent provides information concerning the power connection between the upstream grid and the Microgrid (normal grid operation, grid voltage or frequency imbalances and grid disconnection) as well as the power balance within the Microgrid (total available power generation and demand). Each load or generation node has a graphical representation providing details for its status (ON /OFF) and its operational details (voltage, current, power production or demand, battery’s SOC etc).