Distributed smart grid architecture for delay and price sensitive power management

Inexpensive wireless communication and emerging energy storage technologies are creating a new future for the power grid. These tools are enabling a “smart grid” which can utilize novel control strategies, such as real-time pricing, to achieve efficiency, fairness, and stability. We propose a decentralized smart grid architecture which uses minimal communication to meet these objectives. Then, inspired by queueing models and distributed control in wireless networks, we develop an algorithm for allocating power which is sensitive to both delay as well as price. The delay sensitivity allows for load shifting while the price sensitivity is a natural consumer objective. We show that the algorithm quickly converges even when homes operate asynchronously which makes the architecture practical for implementation. In addition, we show convergence using techniques from lattice theory. Although similar ideas have been used in standard power control algorithms for wireless systems, we demonstrate that the underlying ideas are more general. This provides a theoretical contribution along with a practical contribution to smart grid power management systems.