Robust Passivity Control of Grid-Following Inverters Using Extended-State-Observer-Based Prediction

Passivity theory has emerged as an effective methodology for stability analysis and control design of grid-following inverters (GFL) in weak grids. However, the system’s passivity characteristics become vulnerable when theLCLfilter parameters and active damping (AD) loop gains vary. This paper offers a solution to enhance passivity robustness against these parametric variations. We first develop a multi-loop model that breaks down the system’s admittance into three parts: grid-side current control, capacitor-current AD, and capacitor-voltage AD. The analysis shows that time delays produce negative real parts in the output admittance, which violates passivity conditions. To overcome these limitations, we propose an extended-state-observer (ESO)-based prediction strategy that effectively compensates for delay-induced effects. The method successfully mitigates negative real-part in the output admittance while substantially expanding the passive region and exhibiting improved robustness against parameter mismatches. Experimental results on a lab-constructed test platform validate the proposed approach and are in close agreement with the theoretical analysis.