A Distributed Quasi-Newton Method for Droop-Free Primary Frequency Control in Autonomous Microgrids

There are some technical challenges associated with the traditional droop control method in microgrids, as it is sensitive to measurement errors and is not economical with respect to power sharing. In this paper, a distributed quasi-Newton method for droop-free primary frequency control without central coordination is presented. Based on peer-to-peer communication, each distributed resource exchanges limited information with its neighbors over a sparse communication network. The proposed distributed Newton method, which is a second order method, stabilizes the frequency faster than those widely used first order gradient/subgradient methods. By combining the equal increment rate criteria, the frequency control method minimizes the generation cost and maximizes the utilization of renewable energy in a single process. The simulation results indicate that the proposed method is insensitive to errors and time delays, and can be used to effectively avoid overshooting. Compared to conventional droop control schemes, this method has much better dynamic performance and the costs of power generation are lower.