As renewable energy resources and distributed generation become central to grid operations, the reliability of inverter-based resources (IBRs) faces new challenges. Modern grids depend heavily on inverters to convert DC power from solar panels and batteries into AC electricity while simultaneously providing voltage support, frequency regulation, and other grid services. When hardware components degrade—such as power semiconductor switches or output filters—conventional control strategies often fail to respond effectively.
Traditional fault-tolerant controllers require rapid, accurate fault detection before they can reconfigure, but gradual actuator wear produces ambiguous signals that may never trigger a clear alarm. Adaptive control systems, designed to adjust dynamically, assume the inverter retains enough control authority to track changing setpoints; when actuators degrade, this assumption breaks down, causing the controller to misinterpret tracking errors and drift into unsafe parameter ranges.
The proposed human-on-the-loop (HOTL) architecture addresses these limitations by embedding operator judgment directly into the control feedback loop. Rather than relying solely on algorithms, the system surfaces subtle performance anomalies to human supervisors, who can validate controller decisions, override parameters when warranted, and deliberately relax performance targets if conditions exceed the system's modeled fault space.
Two new metrics enable this resilient decision-making. Generation Reserve Capacity (GRC) quantifies how much additional power or reactive support the inverter can still provide before saturation—essentially measuring operational headroom remaining after degradation. Controlled Performance Degradation (CPD) allows temporary, intentional relaxation of voltage tracking accuracy to preserve overall grid stability when demands exceed the inverter's remaining capacity.
Simulation results demonstrate that the approach prevents actuator saturation, maintains reserve capacity for future grid events, and achieves better voltage regulation than either conventional adaptive control or passive fault-tolerant methods alone. The framework scales to address sequential degradation events—realistic in aging infrastructure—and provides operators with transparent, actionable information for managing aging or damaged resources without losing their grid-support contributions entirely.
