Distributed event‐triggered algorithms for the management of networked battery systems

Abstract

Distributed algorithms have recently been developed for a networked battery system composed of multiple battery units, each of which serves as an energy storage unit. These emerging distributed algorithms rely on continuous communication among battery units and/or continuous update of power allocation for each battery unit. To reduce the communication and computational burden, we propose distributed event-triggered algorithms for the management of networked battery systems with unknown battery unit parameters. In particular, two event-triggered distributed estimators are designed for each battery unit to respectively estimate the average battery unit state and the average desired power with any specified level of accuracy. Two dynamic event-triggering mechanisms are proposed for the distributed estimators that determine when each battery unit communicates with its neighbors. Based on these two event-triggered distributed estimators, an event-triggered distributed adaptive power allocating law along with a static event-triggering mechanism is designed for each battery unit to update the power allocation in a discrete-time manner. It is shown that the proposed power allocating laws achieve state-of-charge balancing among all battery units while delivering the desired total power in either the charging or discharging mode without requiring either continuous communication or continuous update of power allocation. It is also shown that the Zeno behavior is excluded for all the three event-triggering mechanisms. A simulation example is provided to validate the effectiveness of the proposed design.