A Deep Reinforcement Learning-Based Intelligent Grid-Forming Inverter for Inertia Synthesis by Impedance Emulation

In this letter, impedance emulation is exploited for synthesizing inertia in autonomous microgrids. An intelligent grid-forming inverter (GFI) is proposed that facilitates sufficient degrees of freedom for adaptive impedance shaping. The latter adaptively changes the effective bandwidth of the inverter's voltage controller, in response to disturbances for inertia synthesis. Deep reinforcement learning is utilized to tackle the lack of explicit quantitative relation between impedance shaping and inertia. Simulation results prove the effectiveness of the method.

Automated summary

This letter presents a method of using impedance emulation to synthesize inertia in autonomous microgrids. An intelligent grid-forming inverter (GFI) is proposed to provide sufficient degrees of freedom for adaptive impedance shaping. The latter adaptively changes the effective bandwidth of the inverter's voltage controller in response to disturbances for inertia synthesis. Deep reinforcement learning is used to overcome the lack of explicit quantitative relationship between impedance shaping and inertia. Simulation results demonstrate the effectiveness of the method.