🎓 Lesson 15
D5
Droop vs. Virtual Oscillator Control: Stability Implications
Droop control makes a grid-forming converter act like a traditional power plant by reducing its output frequency when it supplies more power, while virtual oscillator control makes it behave like a synchronized physical oscillator that naturally adjusts to grid conditions without needing a frequency reference.
🎯 Learning Objectives
- ✓ Analyze small-signal stability margins (e.g., eigenvalues, damping ratio) for droop-controlled vs. VOC-based grid-forming converters under varying grid stiffness
- ✓ Explain the physical interpretation of virtual inertia and damping coefficients in VOC and their impact on transient frequency nadir and RoCoF
- ✓ Design droop gain (R_droop) and VOC parameters (α, β, ω₀) to meet IEEE 1547-2018 and IEC 62738 stability requirements for weak-grid applications
- ✓ Apply time-domain simulation (e.g., MATLAB/Simulink or PSCAD) to compare frequency recovery performance after a 3-phase fault for both controls
📖 Why This Matters
As mines increasingly deploy solar-diesel-battery microgrids for remote operations—and rely on grid-forming inverters to black-start critical loads—choosing between droop and virtual oscillator control isn’t academic: it determines whether your power system survives a sudden load rejection or generator trip. Droop is widely deployed but can destabilize under ultra-weak grids (SCR < 2); VOC offers superior synchronization resilience but demands deeper understanding of nonlinear dynamics. This lesson equips you to select, tune, and validate the right control for mine site reliability.
📘 Core Principles
Droop control relies on linearized approximations of synchronous machine behavior: it injects artificial 'stiffness' via static P–f and Q–V relationships, making it intuitive and field-proven—but inherently lacks inertia and suffers from steady-state errors and poor transient synchronism in low-inertia systems. In contrast, VOC derives from port-Hamiltonian modeling: it defines a virtual energy storage function (Hamiltonian) and shapes dissipation (damping) and coupling (synchronizing torque) terms to replicate the swing equation’s geometric structure. This yields intrinsic angle-locking, adaptive inertia, and provable asymptotic stability—even for heterogeneous converter fleets—without communication or centralized coordination. Critically, VOC’s stability is governed by Lyapunov-based conditions, not gain tuning alone.
📐 Key Stability Metrics & Parameter Design
The dominant stability metric for both controls is the closed-loop eigenvalue location in the s-plane. For droop, the key design parameter is R_droop (droop gain); for VOC, it’s the virtual impedance coefficient α and damping coefficient β. The natural frequency ω₀ sets nominal output frequency and must match grid nominal (e.g., 50/60 Hz).
💡 Worked Example
Problem: A mine microgrid uses a 2 MW VOC-based battery inverter feeding a weak AC network (SCR = 1.8). Given α = 0.05 pu, β = 0.15 pu, ω₀ = 2π×50 rad/s, and line impedance Z_line = 0.15 + j0.45 Ω, calculate the dominant eigenvalue real part (σ) and assess damping using the linearized VOC model σ ≈ −β/2 − α·Re(Y_grid)/2.
1.
Step 1: Compute grid admittance Y_grid = 1/Z_line = 1/(0.15 + j0.45) ≈ 0.6 − j1.8 S → Re(Y_grid) = 0.6 S
2.
Step 2: Normalize α and β to per-unit (base S_base = 2 MVA, V_base = 400 V): α_pu = 0.05, β_pu = 0.15
3.
Step 3: Apply stability approximation: σ ≈ −0.15/2 − 0.05×0.6/2 = −0.075 − 0.015 = −0.090 pu (≈ −28.3 rad/s in absolute units)
4.
Step 4: Compare |σ| > 0 and σ < 0 → stable; damping ratio ζ ≈ |σ| / ω_n, where ω_n ≈ √(α·Re(Y_grid)) ≈ √(0.05×0.6) ≈ 0.173 → ζ ≈ 0.090 / 0.173 ≈ 0.52 (well-damped)
Answer:
The dominant eigenvalue has real part σ = −0.090 pu, indicating stable, well-damped response (ζ ≈ 0.52), satisfying IEEE 1547-2018 requirement for ζ ≥ 0.3 under weak-grid conditions.
🏗️ Real-World Application
At the Agnew Gold Mine (Western Australia), a 15 MW solar + 10 MW/20 MWh battery microgrid uses VOC-based inverters for black-start and islanded operation. During commissioning, droop-controlled inverters exhibited sustained oscillations (>2 Hz) following a 5 MW load step—requiring manual gain retuning and limiting minimum SCR to 2.5. After switching to VOC with α = 0.035 pu and β = 0.12 pu, the system achieved stable operation at SCR = 1.6, passed EN 50160 voltage flicker compliance, and reduced frequency nadir from 48.2 Hz to 49.1 Hz during diesel trip tests—directly enabling uninterrupted ventilation fan operation.