Abstract
Inverter-connected resources can improve transient stability in low-inertia grids by injecting active power to minimize system frequency deviations following disturbances. In practice, most generation and load disturbances are step changes and the engineering figure-of-merit is often the peak overshoot in frequency resulting from these step disturbances. In addition, the inverter-connected resources tend to saturate much more easily than conventional synchronous machines. However, despite these challenges, standard controller designs must deal with averaged quantities through $H_2$ or $H_\infty$ norms and must account for saturation in ad hoc manners. In this paper, we address these challenges by explicitly considering $L_\infty$ control with saturation using a linear matrix inequality-based approach. We show that this approach leads to significant improvements in stability performance.
Introduction: The Challenge of Low-Inertia Grids
As modern power systems transition from massive, spinning synchronous generators to renewable resources like solar panels and wind turbines, they lose "mechanical inertia." In traditional grids, this inertia acts as a shock absorber; however, in low-inertia microgrids, frequency can spike or plummet rapidly following load changes.
While power inverters can react nearly instantly to stabilize the grid, they are subject to strict actuator saturation limits. Unlike traditional generators, an inverter has a relatively low power ceiling. If a controller requests power beyond this limit, the system becomes nonlinear and potentially unstable. Most current control designs optimize for "average" performance (using $H_2$ or $H_\infty$ norms), which fails to capture the most critical factor for grid stability: the peak overshoot. If frequency hits a specific threshold even momentarily, protection relays trip, potentially leading to a blackout.
Methodology: $L_\infty$ Optimization and Invariant Ellipsoids
To address these issues, this research proposes a controller synthesis approach based on Linear Matrix Inequalities (LMIs) and the method of invariant ellipsoids.
Key components of the methodology include:
- $L_\infty$ Optimization: Instead of minimizing average error, the design minimizes the maximum possible deviation (the $L_\infty$ norm) resulting from bounded disturbances.
- Invariant Ellipsoids: These geometric shapes are used to "trap" the system's state. If the system starts within the ellipsoid and disturbances remain within a defined range, the system is guaranteed to stay within the safe bounds of the ellipsoid.
- The *-norm: Since exact $L_\infty$ calculation is computationally intensive, the researchers utilize the "*-norm" as a tractable upper bound.
- High-Gain Control: The design intentionally pushes the inverter to its saturation limits. Because the approach is rooted in invariant ellipsoids, the system remains stable even when the inverter is operating at its power limit (effectively a "bang-bang" control style).
Key Results: Superior Peak Mitigation
The controller was evaluated using a simulated single-area power system model consisting of an aggregate generator, a turbine governor, and an inverter-connected resource. The system was subjected to synthetic disturbances comprising random step changes in load.
The simulations revealed several key insights:
- Superior Overshoot Mitigation: The proposed $L_\infty$ controller significantly reduced frequency deviations compared to standard Linear Quadratic Regulator (LQR) and pole-placement designs.
- Efficiency of High-Gain: The "high-gain" version of the controller outperformed the "low-gain" linear version by utilizing the full power capacity of the inverter more frequently, as seen in the control input saturation in Figure 6(b).
- Tight Performance Bounds: The invariant ellipsoids provided a reasonable overbound of actual system behavior, validating the LMI-based optimization as an effective tool for real-world power constraints.
Conclusion and Future Directions
This work demonstrates that $L_\infty$ control is fundamentally better suited for frequency regulation in power systems than common $H_2$ or $H_\infty$ methods because it prioritizes peak overshoot. By treating saturation as a known constraint rather than a bug, the controller can safely "ride the limit" to provide maximum stabilization.
While this study utilized a linearized, single-area model and ignored internal inverter dynamics, it sets the stage for future research. Subsequent work will focus on extending this framework to multi-node systems where multiple controllable inverters must coordinate to stabilize complex, large-scale grids.
Summary updated