⚡ electrical engineering

Decentralized Small Gain and Phase Stability Conditions for Grid-Forming Converters: Limitations and Extensions

This paper proposes a novel decentralized stability framework for grid-forming converters that overcomes the limitations of existing small gain and phase criteria by introducing loop shaping transformations to address low-frequency non-sectoriality, thereby reducing conservativeness and enabling scalable stability certification in power systems.

Original authors: Diego Cifelli, Adolfo Anta

Published 2026-04-15

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Diego Cifelli, Adolfo Anta

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Big Picture: The "Dance Floor" of the Power Grid

Imagine the modern power grid as a massive, high-energy dance floor. In the past, this floor was dominated by heavy, steady dancers (traditional power plants like coal or nuclear) who moved in a predictable, rhythmic way.

Today, the floor is getting crowded with new dancers: Solar panels, wind turbines, and battery storage. These are "converter-based" resources. They are fast, agile, and use complex electronics to move. The problem? When too many of these new dancers try to move in sync with the old rhythm, they sometimes trip over each other, causing the whole floor to shake or collapse (blackouts).

Engineers need a way to check if the dance floor is stable before the music starts. Traditionally, they did this by simulating the entire dance floor on a supercomputer, trying every possible combination of dancers. But as the grid grows to hundreds of converters, this simulation becomes too slow and expensive.

The Old Solution: The "Perfectly Polite" Rule

To avoid simulating everyone, researchers invented a decentralized rule. Think of it like a bouncer at the door. Instead of watching the whole party, the bouncer checks each dancer individually.

The Rule: "If every single dancer is polite (passive) and doesn't move too wildly (small gain), the whole party will be safe."

This works great for the old heavy dancers. But for the new electronic converters, it fails for two reasons:

They aren't "polite" at low speeds: At low frequencies (slow movements), these converters act like constant-power machines. They try to push or pull with a fixed force regardless of the voltage, which breaks the "politeness" rule.
The rule is too strict: Even when the converters are behaving, the old rule is so conservative (scared of risk) that it says "Danger!" even when there is no danger. It's like a bouncer kicking out a dancer just because they are wearing a slightly loud shirt, even though they aren't actually dancing badly.

The Problem: The "Sectoriality" Wall

The paper identifies a specific mathematical wall called Sectoriality.
Imagine trying to draw a line on a map to separate "Safe" from "Unsafe."

For the new converters, their behavior at low frequencies is so chaotic that you cannot draw a straight line to separate them. They exist in a "gray zone" where the old math says, "I can't tell if this is safe or not."
Because the math can't define a safe zone, the decentralized check fails, and engineers are forced to go back to the slow, expensive simulations.

The New Solution: Changing the Camera Angle

The authors of this paper realized the problem wasn't that the converters were "unsafe," but that we were looking at them through the wrong lens.

The Analogy: The Spinning Top
Imagine a spinning top.

If you look at it from the side (the standard "rectangular" view), it looks like it's wobbling wildly and might fall over.
But if you look at it from directly above (the "polar" view), you see it spinning perfectly in a circle. It's stable!

The paper proposes a Loop-Shaping Transformation. This is like putting on special 3D glasses or rotating the camera.

At Low Frequencies (Slow movements): They switch the view to "Power-Polar coordinates." In this view, the converter looks like a stable, spinning top. The "gray zone" disappears, and the math can finally say, "Yes, this is safe."
At High Frequencies (Fast movements): They switch back to the standard view, which works fine for fast electronics.

The "Magic Filter": Making it Work Everywhere

The tricky part is that you can't just switch cameras instantly; the transition would look jerky and break the math.

The authors created a smart, frequency-dependent filter.

Think of this as a chameleon suit.
When the music is slow (low frequency), the suit turns into the "Polar View" to show stability.
When the music speeds up (high frequency), the suit smoothly morphs into the "Standard View."
This ensures the converter looks stable and well-behaved across the entire speed range, not just in one spot.

Why This Matters: The "Virtual Admittance" Trick

The paper also introduces a clever trick to make the safety check even less scary (less conservative).

Imagine two dancers holding a spring between them. If the spring is stiff, they might push against each other too hard.
The authors suggest: "Let's pretend the spring belongs to the other dancer."
By mathematically shifting the "virtual spring" (virtual admittance) from the converter to the network, they can tighten the safety bounds. This allows them to prove the system is stable even when the converters are pushing harder, without actually changing the hardware.

The Results: Testing the Theory

The authors tested this on two scenarios:

A Single Converter: They showed that with their new "camera angle," a converter that looked unstable under the old rules was proven stable.
The IEEE 14-Bus System: This is a complex map of a real power grid with 14 nodes and 4 converters.
- They created a scenario where the system should be unstable (one converter was "drunk" with low damping).
- Their method successfully identified exactly which converter was causing the trouble (Converter #8).
- By fixing just that one converter, the whole system became stable again.

The Bottom Line

This paper gives engineers a scalable, decentralized toolkit.

Before: We had to simulate the whole grid to be sure it was safe, or we had to assume the worst-case scenario and restrict how much renewable energy we could add.
Now: We can check each converter individually using a "smart camera" that knows how to look at them correctly. This proves stability more accurately, reduces unnecessary restrictions, and paves the way for a greener, more converter-heavy power grid without the fear of blackouts.

In short: They didn't fix the dancers; they just taught us how to watch them so we can stop worrying about them tripping.

1. Problem Statement

As power systems transition toward converter-dominated grids, ensuring stability without exhaustive system-wide simulations is critical. Decentralized stability criteria, specifically Mixed Small Gain and Phase conditions, have been proposed to certify stability by verifying local conditions on individual converters and the network.

However, the application of these criteria to Grid-Forming (GFM) converters faces three major limitations:

Violation of Small-Gain: GFM converters often exhibit high low-frequency gain (due to constant-power characteristics), causing the converter gain to exceed the network gain, violating the small-gain condition.
Violation of Sectoriality: The small-phase criterion requires system admittances to be "sectorial" (their numerical range must lie within an angular sector not containing the origin). At low frequencies, GFM admittances are intrinsically non-sectorial due to their constant-power nature, rendering the phase criterion inapplicable.
Conservatism: Even when sectoriality holds, the standard criteria often yield overly conservative bounds, incorrectly predicting instability in stable systems due to worst-case phase pairing assumptions.

2. Methodology

The authors propose a generalized framework that extends the decentralized mixed gain-phase theory through Loop-Shaping Transformations.

A. Analysis of Non-Sectoriality

The paper first mathematically proves that the non-sectoriality of GFM converters at low frequencies is an intrinsic property resulting from their constant-power behavior (voltage-to-current relationship), rather than a specific control implementation flaw. This holds true for both Grid-Forming and Grid-Following (GFL) converters with standard synchronization loops.

B. Power-Polar Coordinate Transformation

To resolve the low-frequency non-sectoriality, the authors introduce a change of variables from the standard rectangular ( $d-q$ ) frame to Power-Polar coordinates (Voltage Magnitude/Phase as inputs; Active/Reactive Power as outputs).

In this frame, GFM converters exhibit passivity and sectoriality at low frequencies.
This is achieved via a linear transformation matrix ( $E, F, C$ ) applied to the converter and network admittances, effectively reshaping the loop.

C. Unified Frequency-Dependent Approach

Since the Power-Polar frame is effective at low frequencies but the standard rectangular frame is better at high frequencies, the authors propose a unified, frequency-dependent transformation:

Filtering Strategy: They use low-pass and high-pass filters to blend the transformation matrices. At low frequencies, the system uses the Power-Polar transformation; at high frequencies, it reverts to the identity (rectangular) transformation.
Weighting Matrix ( $W_i$ ): An additional degree of freedom is introduced via a weighting matrix $W_i(s)$ . This allows for Virtual Admittance Compensation, where known dynamics (e.g., virtual impedance) are shifted from the converter model to the network model. This tightens the phase bounds and reduces conservatism.

3. Key Contributions

Identification of Intrinsic Limitations: The paper rigorously demonstrates that the failure of standard decentralized criteria for GFM converters is due to intrinsic constant-power dynamics, not just poor control tuning.
Loop-Shaping Extension: It extends the decentralized small-phase theorem by introducing a loop-shaping transformation that allows the system to be analyzed in alternative coordinate frames (Power-Polar) where sectoriality is satisfied.
Unified Framework: It proposes a continuous, frequency-dependent transition between coordinate frames, ensuring stability certification across the entire frequency spectrum.
Conservatism Reduction: By utilizing the weighting matrix $W_i(s)$ , the method allows for "virtual admittance" compensation, significantly tightening the phase bounds and avoiding false instability alarms.

4. Results and Validation

The methodology was validated through two case studies:

Case 1: Infinite Bus System:
- Standard Approach: Failed to certify stability due to non-sectoriality below 0.9 Hz and high conservatism between 0.9 Hz and 50 Hz.
- Proposed Approach: The frequency-dependent transformation restored sectoriality across the full frequency range.
- Virtual Admittance: Applying the $W_i(s)$ compensation (mimicking virtual impedance) further reduced phase spread, allowing stability certification even in weak grid conditions.
- Unstable Case: The method successfully detected instability when GFM damping was reduced, showing a clear violation of the phase condition at the oscillation frequency (1.2 Hz).
Case 2: IEEE 14-Bus System:
- A multi-converter system (4 GFMs) was analyzed.
- The decentralized conditions successfully identified the specific converter (Bus 8) responsible for instability when its damping was lowered.
- The method proved robust even when the exact virtual admittance parameters were unknown, using approximate values for the weighting matrix to achieve effective stability validation.

5. Significance

Scalability: The approach enables stability certification for large-scale converter-dominated grids without requiring full-system simulations, as it relies on local converter models and the network admittance.
Grid-Strength Independence: The small-phase condition, when combined with the proposed transformation, provides stability guarantees that are largely independent of grid strength (Short Circuit Ratio), a significant improvement over gain-based methods.
Practical Application: The method offers a pathway for updating grid codes, allowing for less conservative stability certificates that can accommodate the diverse dynamics of modern Grid-Forming converters.
Future Outlook: While highly effective for GFMs, the authors note that further work is needed to integrate Grid-Following converters and synchronous machines, particularly regarding auxiliary control loops (AVR, PSS).

In summary, this paper resolves the fundamental applicability gap of decentralized small-phase criteria for GFM converters by mathematically proving the source of their limitations and providing a robust, transform-based solution that ensures stability certification across all frequencies.