The Cooperative Vehicular Networking Laboratory of the School of Artificial Intelligence recently presented the research progress, themed An Event-Triggered Deadbeat Control Considering Dynamic Power Loss Compensation for Hybrid Energy Storage System in the top journal IEEE Transactions on Industrial Electronics (IF: 8.162).

Accelerating the construction of a modern energy system is a solution for ensuring national energy security and achieving "carbon peaking" and "carbon neutrality" as scheduled. The modern power system is required to be not only reliable but also more environmentally friendly and efficient. The current generation system, electrical transmission network, and distribution network needs to be upgraded to meet all the needs. To achieve it, the concept of the smart grid is proposed as the solution. Compared with the traditional electricity network, the distributed network of the smart grid is more active and integrates more renewable energy sources. All actions of users and generators can be integrated together, in order to efficiently deliver sustainable, economical, and reliable supplies through the advanced control strategy and communication technology.

The scheme of event-triggered deadbeat control

Renewable energy sources cannot work out for a long time because the power generated by renewable energy sources is intermittent and is affected by environmental conditions easily. Hence, to regulate the DC bus voltage as a constant value, a hybrid energy storage system (HESS) is used to compensate for the intermittent power.

The hardware setup.

This paper proposed an event-triggered deadbeat control for HESS to regulate the bus voltage of microgrids efficiently. The charging and discharging behaviors of HESS are controlled by the deadbeat-based controller, and the turn-on resistance of the circuit is considered in the system modeling, to eliminate the steady-state error caused by the conduction losses. Under the control of the event-triggered mechanism, the switching actions for DC-DC converters would be triggered only if the variation of state variables exceeds the safe range, which can reduce switching losses to reach a better performance of regulation, and the computational burden can be reduced.

The experimental results

The simulation and hardware experiments substantiate that there is no state-steady error under the control of the proposed method, and the dynamic error and settling time are reduced by 25.6%-36% and 33.3%-37.5%, compared with conventional deadbeat control. Under the control of the event-triggered mechanism, the switching losses and computational burden are reduced by 30%-41.9%.