Hardware test and validation of the angular droop control: Analysis and experiments

This paper presents a hardware-based validation of angular droop control for grid-forming DC/AC converters, demonstrating its ability to unify primary and secondary control for exact frequency regulation while addressing implementation challenges like discretization and clock drift through single and multi-converter experiments.

The Gist

Imagine the electrical grid as a massive, global orchestra. For over a century, this orchestra has been conducted by giant, spinning machines (synchronous generators) that naturally keep everyone in time. But today, we are replacing those heavy machines with fast, silent, and flexible digital musicians (solar panels, wind turbines, and batteries).

The problem? Digital musicians don't have the same "natural rhythm" as the spinning machines. If they don't talk to each other perfectly, the music falls apart, and the lights go out.

This paper is about teaching these digital musicians a new, smarter way to keep the beat. The authors built a real-life hardware test lab to prove that a new control method called "Angular Droop" works better than the old methods.

Here is the breakdown in simple terms:

1. The Old Way vs. The New Way

The Old Way (Frequency Droop):
Imagine a group of runners trying to stay in sync. If one runner slows down because they are tired (a change in power load), the old rule says: "If you slow down, the whole group must slow down a tiny bit to match you."

• The Result: The group's speed (frequency) drifts away from the perfect pace. To fix it, a coach (a secondary control layer) has to run over and yell, "Speed up!" to get them back on track. It's a two-step process: react, then correct.

The New Way (Angular Droop):
The new method is like a group of dancers who know exactly where they are on the stage (the "angle") relative to the music.

• The Magic: Instead of reacting to how fast they are moving, they react to where they are. If the music gets louder (more power needed), they instantly adjust their position to match the beat perfectly.
• The Benefit: They never drift off-beat in the first place. They combine the "reacting" and the "correcting" into one single, instant move. The frequency stays perfect automatically.

2. The Hardware Challenge: The "Drifting Clock" Problem

The authors didn't just simulate this on a computer; they built it with real wires, switches, and lights. This is where things got tricky.

The Problem:
Imagine two musicians trying to play a duet. If Musician A's watch is 1 second fast and Musician B's watch is 1 second slow, they will eventually drift out of sync, even if they try their best. In the digital world, this is called Clock Drift. Since "Angular Droop" relies on counting time to know the "angle," even a tiny drift in the computer's clock causes the system to fail.

The Solution:
The team solved this by giving all their computers a Master Clock. Think of it like a conductor tapping a baton that sends a precise "tick-tock" signal through a fiber-optic cable to every musician. Now, everyone shares the exact same time, eliminating the drift.

3. The Experiments: Proving it Works

They tested their system in two scenarios:

Scenario A: The Soloist (Black Start)

• The Test: They turned the power off completely (a "blackout") and then tried to turn it back on.
• The Result: The system successfully "woke up," created a perfect wave of electricity from nothing, and handled sudden changes in load (like flipping on a heavy heater) without losing its rhythm. It proved the system can start a grid from scratch.

Scenario B: The Duet (Power Sharing)

• The Test: They connected two converters to the same load, like two generators sharing the work of powering a factory.
• The Result: The two units instantly synchronized their rhythm and split the workload evenly. If one tried to do too much, the other naturally picked up the slack, all without a central boss telling them what to do.

4. Why This Matters

This paper is a bridge between theory and reality.

• Before: Scientists had math proofs saying "Angular Droop" is great, but no one knew if it would work with real, messy hardware.
• Now: They have proven it works. They showed how to handle the "clock drift" issue and how to tune the system so it's stable.

The Big Picture:
As we move toward a future powered by renewables, we need grids that are self-healing and self-synchronizing. This "Angular Droop" control is like giving every solar panel and battery a built-in sense of rhythm that keeps the entire grid stable, even when the wind stops blowing or the sun goes behind a cloud. It's a crucial step toward a smarter, more resilient energy future.

Technical Summary

Here is a detailed technical summary of the paper "Hardware test and validation of the angular droop control: Analysis and experiments."

1. Problem Statement

The power grid is transitioning from synchronous machines to converter-based renewable energy resources. This shift alters system dynamics, necessitating advanced control strategies for Grid-Forming (GFM) DC/AC converters.

• Limitation of Existing Methods: Traditional frequency droop control (active power-to-frequency droop) relies on the dynamics of synchronous machines. While effective, it results in a steady-state frequency deviation proportional to the power imbalance. This requires a secondary control layer to restore the frequency to its nominal value.
• The Gap: Angular droop control (active power-to-angle droop) theoretically offers exact frequency regulation at steady state by merging primary and secondary control into a single layer. However, previous validations were limited to numerical simulations. Simulations often fail to account for real-world hardware constraints such as discretization effects, clock drifts in distributed systems, and sensor noise.
• Core Challenge: There is a lack of hardware-based validation for angular droop control, specifically addressing the practical challenges of implementing angle-based dynamics in discrete-time digital controllers and synchronizing multiple converters without a global time reference.

2. Methodology

The authors conducted hardware experiments using a programmable DC/AC converter testbed to validate angular droop control in two distinct scenarios.

Experimental Setup:

• Hardware: Two identical 15 kW three-phase DC/AC converters using SiC-MOSFETs, connected to a resistive load via inductive transmission line replicas.
• Control Architecture: A real-time control unit (20 kHz cycle) executes the angular droop algorithm. The system utilizes current and voltage sensors with high bandwidth.
• Scenarios:
  1. Scenario I (Single Converter): A single converter connected to a load to test black-start capabilities, load disturbance rejection, and implementation schemes (Direct vs. Indirect control).
  2. Scenario II (Two Converters): Two converters connected to a common load to test frequency synchronization and power-sharing.

Key Technical Solutions Implemented:

• Discretization of Angle Dynamics: Since angular droop relies on integrating frequency to get an angle, the angle variable grows infinitely over time, causing numerical precision loss in single-precision hardware. The authors proposed a solution mapping the nominal angle to an $n$-dimensional torus ($T^n$) by applying a modulo $2\pi$ operation to the nominal angle reference, while keeping the error dynamics separate. This ensures angles remain within feasible numerical bounds.
• Clock Drift Mitigation: In multi-converter systems, independent local clocks cause "clock drift," leading to frequency mismatches and power oscillations. The authors implemented a master clock distribution system using optical fiber connections between controllers. This synchronizes the local clocks to within $\pm 2$ ns, eliminating drift-induced errors.
• Control Law Reformulation: The control law was adapted for hardware realization, comparing Direct Control (modulation signal derived directly from the angle) and Indirect Control (cascaded voltage/current loops tracking an angle reference).

3. Key Contributions

1. First Hardware Implementation: This paper presents the first experimental validation of angular droop control on a real hardware testbed, bridging the gap between theoretical proofs and practical deployment.
2. Solutions to Implementation Challenges:
   • Proposed a robust discretization method to handle infinite angle growth in digital controllers.
   • Demonstrated a master clock distribution strategy to solve clock drift issues in distributed GFM inverters.
3. Comparative Analysis: Provided a detailed comparison between Direct and Indirect control schemes, showing that while both work, the Direct scheme offers faster convergence and simpler tuning for this specific application.
4. Tuning Guidelines: Established practical guidelines for selecting control gains ($\alpha$ for transient response and $\gamma$ for steady-state power-angle droop) to balance frequency deviation and convergence speed.

4. Experimental Results

Scenario I: Single Converter to Load

• Black Start: The system successfully started from a dead state, forming sinusoidal waveforms at the nominal frequency (50 Hz) and amplitude without external grid support.
• Load Disturbance: Upon a step change in load (from ~2.88 kW to ~3.8 kW), the system demonstrated:
  • Zero Steady-State Frequency Error: Unlike frequency droop, which showed a permanent frequency deviation, the angular droop control returned the frequency exactly to 50 Hz.
  • Fast Transient Response: The system reacted quickly to the load change, with the frequency nadir and settling time dependent on the gain $\alpha$.
• Direct vs. Indirect: The Direct implementation showed faster settling times (0.11s vs 0.24s for load steps) and lower angle errors compared to the Indirect implementation.

Scenario II: Two Converters to Common Load

• Frequency Synchronization: When two converters were connected, they synchronized their frequencies and phases within 0.25s. The master clock distribution successfully prevented frequency drift.
• Power Sharing: By tuning the droop gains ($\gamma_1 = \gamma_2$), the converters achieved equal active power sharing ($P_1 \approx P_2$).
• Reactive Power: Small voltage amplitude mismatches (due to hardware tolerances) resulted in non-zero reactive power exchange, which is expected in inductive networks, but the system remained stable.
• Stability: The experiments validated the theoretical claim that angular droop is an inverse optimal stabilizing control law, satisfying security constraints regarding angle differences ($|\theta_1 - \theta_2| < \pi/2$).

5. Significance

• Elimination of Secondary Control: The results prove that angular droop control can achieve exact frequency regulation at steady state, effectively merging primary and secondary control layers. This simplifies the control architecture for future microgrids and reduces communication overhead.
• Robustness in Real-World Conditions: By addressing discretization and clock drift, the paper demonstrates that advanced angle-based control strategies are viable for industrial deployment, not just theoretical constructs.
• Foundation for Future Grids: The successful validation of black-start capabilities and plug-and-play power sharing in a multi-converter setup provides a strong foundation for the integration of high shares of inverter-based resources in weak or islanded grids.

In conclusion, the paper successfully transitions angular droop control from simulation to hardware, proving its superiority in frequency regulation and stability while providing critical engineering solutions for its real-world implementation.