Scalability in smart grid data flow architectures

The increasing complexity of power networks calls for new decentralised solutions for electrical power system control. Smart grid techniques and technologies are seen as a potential solution, therefore the design of effective distributed data flow architectures are of growing importance. This paper utilises Multi Agent Systems (MAS) to investigate and demonstrate the significance of appropriately designed data flow architectures in performing effective decentralised power system control. Through examining the impact of an increasing control population on control accuracy, communication demand, responsiveness and robustness in the presence of continuous anomalous data; the paper illustrates that data flow architecture is a key design consideration.