PaperlayerView exact revision on arXiv

Source-linked AI summary

Control of Grid-Forming VSCs: A Perspective of Adaptive Fast/Slow Internal Voltage Source

Heng Wu, Xiongfei Wang

arXiv:2212.03053v2eess.SYcs.SY

TL;DR

GFM-VSCs face conflicting internal-voltage dynamics requirements: slow dynamics supports grid-forming services, while fast dynamics helps avoid loss of synchronism. This paper proposes adaptive fast/slow IVS control, and reports experimental effectiveness across grid strengths and disturbance types.

Abstract

from arXiv · show

Grid-forming (GFM) capability requirements are increasingly imposed on grid-connected voltage-source converters (VSCs). Under large grid disturbances, GFM-VSCs need to remain stable while providing GFM services. Yet, such objectives, as pointed out in this paper, inherently lead to conflicting requirements on the dynamics of internal voltage source (IVS) of GFM-VSCs, i.e., the fast IVS dynamics is needed to avoid the loss of synchronism with the grid, whereas the slow IVS dynamics is preferred for maintaining GFM capability. To tackle this challenge, an adaptive fast/slow IVS control is proposed, which switches GFM-VSC between fast and slow IVS dynamics based on system needs. The proposed method enhances the transient stability of GFM-VSC, whilst maximizing its capability of providing GFM service. Further, the approach is robust to different grid strengths and different types of grid disturbances. The experimental results verify the theoretical findings and the effectiveness of the proposed control method.

II. IMPACT OF FAST AND SLOW IVS CONTROL

The paper uses an SLVM-controlled GFM-VSC with adaptive virtual-impedance current limitation as its benchmark, and describes its power, voltage, and current-limiting control paths.

II. IMPACT OF FAST AND SLOW IVS CONTROL

  • The benchmark GFM-VSC connects through an LC filter to the point of common coupling, with grid impedance representing the connection to the grid.The paper treats the converter as a voltage source behind impedance in its equivalent circuit.Source: A. System Description · pp. 2–3
  • Active power control combines P–ω droop with virtual inertia and damping emulation, and outputs the angle reference used for dq transformation.The emulated virtual inertia has the same effect as a low-pass filter, so no additional filter is needed after active-power measurement.Source: A. System Description · p. 3
  • Reactive power control, such as Q–V droop, regulates the converter's voltage-magnitude reference using filtered reactive-power measurement.The paper denotes the voltage reference and nominal voltage magnitude as Vref and Vn, respectively.Source: A. System Description · p. 3
  • The SLVM control regulates output-voltage magnitude to track Vref, while adaptive virtual impedance limits current by adjusting impedance according to fault-current magnitude.The fictitious voltage drop for virtual impedance is subtracted from the voltage reference; filtering its calculation is used to guarantee small-signal stability when virtual impedance is active.Source: A. System Description · p. 3

B. Fast or Slow IVS Control: Performance Comparison

Slow IVS dynamics supports grid-forming services but can jeopardize transient stability during severe disturbances, while fast dynamics limits that risk at the cost of weaker disturbance response. The proposed adaptive control switches between them according to operating conditions.

B. Fast or Slow IVS Control: Performance Comparison

  • For a small grid phase jump, slow IVS leaves the internal voltage nearly unchanged and increases output active power, whereas fast IVS limits the power response.The additional active power from slow IVS is described as active phase jump power; slow IVS can also provide fast inertia power during frequency drifts.Source: B. Fast or Slow IVS Control:… · p. 3, Figure 3 · p. 4
  • Under a large phase jump, slow IVS can drive the power angle past the unstable equilibrium and cause loss of synchronism, while fast IVS limits the angle change and avoids it.The paper states that the same conclusions apply to grid voltage dips.Source: B. Fast or Slow IVS Control:… · p. 4, Figure 4 · p. 4
  • Slow IVS supports grid-forming capability but can cause loss of synchronism under severe disturbances; fast IVS has the opposite effect.The comparison identifies the trade-off between capability and transient stability.Source: Table I · p. 4, B. Fast or Slow IVS Control:… · p. 4

A. Basic Idea

  • The adaptive scheme uses fast IVS at the current limit and, below that limit, uses slow IVS at small power angles but fast IVS at large angles.Because estimating the power angle requires grid voltage and impedance, the paper identifies a local-measurement switching criterion as the practical need.Source: A. Basic Idea · p. 5, Figure 5 · p. 5

B. Iomag-based Switching Criterion

Before current limiting, output-current magnitude rises monotonically with power angle, enabling a local-current switching criterion. A worst-case threshold provides a sufficient condition for keeping the angle below its selected limit and avoiding current limitation.

B. Iomag-based Switching Criterion

  • Iomag rises monotonically with δ before current limitation, so Iomag < Iomag_th_min is a sufficient condition for δ < δth.The threshold is calculated for the largest considered grid impedance and a selected grid-voltage condition.Source: B. Iomag-based Switching Criterion · p. 5
  • With Xf = 0.1 p.u., Xg = 1 p.u., and Eref = 1.1 p.u., Iomag_th_min peaks around 1 p.u., below the selected 1.5 p.u. current limit.Thus, Iomag < Iomag_th_min also ensures Iomag < Ilim under the stated setup.Source: B. Iomag-based Switching Criterion · p. 6, Figure 6 · p. 6
  • Choosing δth below 90° trades transient-stability margin against GFM capability; the paper selects δth = 70°, yielding Iomag_th_min = 0.94 p.u.The 90° boundary corresponds to the stated worst case of SCR = 1.Source: B. Iomag-based Switching Criterion · p. 6

C. Discussion and Practical Realization

The practical criterion uses output-current magnitude to switch between slow and fast IVS control, with hysteresis and delay to limit noise-driven switching. The fast-IVS approach modifies APC behavior to retain the power angle during disturbances.

C. Discussion and Practical Realization

  • The 0.94 p.u. current threshold corresponds to δ = 70° only in the worst case SCR = 1; at SCR = 10, it corresponds to δ = 10°.At SCR = 10, current limitation also occurs at a small angle: Iomag = 1.5 p.u. at δ = 15°.Source: C. Discussion and Practical Realization · p. 6
  • The worst-case current threshold can switch to fast IVS at a small angle in a stiff grid, slightly reducing GFM capability versus an ideal angle-based criterion.The passage says the degradation is not significant considering current-limit control.Source: C. Discussion and Practical Realization · p. 6
  • The practical switch uses hysteresis and a time delay: fast IVS starts above Iomag_th_min, while slow IVS resumes below 0.9Iomag_th_min for t1 = 0.2 s.A longer t1 improves noise immunity but reduces sensitivity.Source: C. Discussion and Practical Realization · p. 6

A. Basic Idea

  • Fast IVS control modifies APC to retain δ by making θref follow θg during disturbances; the paper adds an Iomag-based Pref adjustment to address limitations of existing approaches.The approaches discussed modify θref through ω or Pref.Source: A. Basic Idea · pp. 6–7

B. Hybrid-Synchronization Control

The paper revisits two transient-stability controls as fast-IVS mechanisms and shows their effectiveness depends on disturbance type and grid strength. Their complementary strengths do not remove their shared degradation under weak-grid conditions.

B. Hybrid-Synchronization Control

  • HSC uses Voq feedforward into ω to counter an increasing δ, but its effect depends on (1-a) and Vg.The feedforward is more effective for phase-angle jumps and stiff grids, and less effective for voltage dips and weak grids.Source: B. Hybrid-Synchronization Control · p. 7, Figure 9 · p. 7

C. Vomag-based Pref Adjustment

D. Pref-Iomag Droop Control

The proposed Pref–Iomag droop joins other fast-IVS controls on top of standard slow-IVS control and is switched according to the current-based criterion. The reported analysis finds stable poles with the selected parameters, and experiments demonstrate transient-stability effectiveness in stiff-grid disturbances.

D. Pref-Iomag Droop Control

  • For any given grid strength, Iomag increases with δ, making Pref reduction based on Iomag a fast-IVS control more robust to grid-strength variations.The control is named Pref–Iomag droop.Source: D. Pref-Iomag Droop Control · p. 8
  • The droop activates above IPth; with n = 10 p.u., a 0.1 p.u. Iomag increase reduces Pref by 1 p.u., while a limiter keeps Pref1 between 0 and 1 p.u.The selected IPth is 1.1 p.u.; the droop coefficient balances stability benefit against noise-driven Pref swings.Source: D. Pref-Iomag Droop Control · p. 8
  • Fast IVS combines Pref–Iomag droop, Vomag-based Pref adjustment, and Voq feedforward atop slow IVS, switching according to the defined criterion.When fast IVS is off, the system returns to standard slow-IVS control, which guarantees GFM capability.Source: D. Pref-Iomag Droop Control · p. 8, Figure 11 · p. 8, Figure 12 · p. 8

E. Small-Signal Stability Analysis

V.EXPERIMENTAL RESULTS

  • Experimental tests were conducted to validate the proposed adaptive fast/slow IVS control for the GFM-VSC, with controller parameters tuned to guarantee small-signal stability.The setup used a 45 kVA Chroma 61850 grid simulator to generate grid voltage and emulate different grid disturbances; the algorithm ran on a DS1007 dSPACE system.Source: V.EXPERIMENTAL RESULTS · p. 8

A. Verification of Switching Criterion

The current-magnitude switching criterion activates fast IVS control during a severe fault and returns to slow control after a delay. Under undisturbed conditions, adaptive control retains the slow dynamics that support GFM services during RoCoF and phase-angle events.

A. Verification of Switching Criterion

  • During a 0.1 p.u. voltage fault, current magnitude rises to 1.5 p.u. and activates fast IVS control; it switches back to slow control 0.2 s after fault clearance.The measured current falls below 0.9 times the minimum threshold before the delayed return.Source: A. Verification of Switching Criterion · p. 9, Figure 15 · p. 9

B. Verification of GFM Capability

  • During a -5 Hz/s RoCoF event, adaptive control retains slow IVS dynamics and provides active droop and additional inertia power, unlike always-fast control, which degrades inertial response.The test used a 0.1 s RoCoF event at SCR=10 with zero initial output power.Source: B. Verification of GFM Capability · p. 9, Figure 16 · p. 10
  • For a -10° phase-angle jump, adaptive control retains slow IVS dynamics and active phase-jump power, whereas always-fast control produces much less phase-jump power.The comparison was conducted without triggering current limitation.Source: B. Verification of GFM Capability · p. 9, Figure 17 · p. 10

C. Verification of Transient-Stability Improvement

Experiments show that adaptive fast/slow IVS control improves transient stability across stiff and weak grids and several disturbance types. Under weak-grid conditions, the Pref-Iomag droop control is important for stabilization; tests also cover asymmetrical faults and compare against an existing method.

Stiff grid

Weak grid

Asymmetrical fault-ride through

  • Adaptive fast/slow IVS control rides through single- and double-line-to-ground faults in both stiff and weak grids; the paper attributes the comparatively favorable transient-stability case to positive-sequence voltage remaining at 0.33 p.u. even for severe asymmetrical faults.The comparison is with symmetrical faults, where remaining positive-sequence voltage could be below 0.33 p.u.Source: C. Verification of Transient-Stability… · p. 15, Figure 26 · p. 16, Figure 27 · p. 16

Compared with other state-of-the-art method

  • In experimental comparisons under large grid disturbances with current limiting triggered, the proposed adaptive fast/slow IVS control stably rides through, whereas the control exhibits loss of synchronism.The method enhances damping by frequency feedforward in the RPC loop, which is naturally bypassed under current limiting.Source: C. Verification of Transient-Stability… · p. 16

VI.CONCLUSION

The paper proposes adaptive fast/slow IVS control to address the differing IVS dynamics needed for transient stability and GFM capability. Experiments support stability across grid strengths and disturbance types, and a comparison shows stability where the method in [18] becomes unstable under current limiting.

Conclusion

  • Adaptive fast/slow IVS control is proposed to address the distinct IVS dynamic requirements for transient stability and GFM capability.The method is reported to retain GFM capability when stability and current margins are sufficient.Source: VI.CONCLUSION · p. 16
  • Experiments verify transient stability across grid strengths and voltage dips, phase-angle jumps, high RoCoF, and their combinations.The conclusion also reports retained GFM capability when stability and current margins are sufficient.Source: VI.CONCLUSION · p. 16
  • Under a -60° phase jump and -5 Hz/s RoCoF, the proposed control is stable while the method in [18] is unstable.The comparison is shown for SCR=10 and Pm=1 p.u.Source: Figure 28 · p. 17

APPENDIX

The appendix distinguishes the controller-dq and system-dq frames and develops small-signal models for control components. It combines component models into a complete fast-IVS model; during disturbances, saturation can bypass the RPC loop and SLVM control.

Appendix

  • The controller-dq frame is defined by the APC output phase angle, while the system-dq frame aligns with grid-voltage phase angle; superscripts c and s denote their variables.The phase-angle difference δ is defined as θref - θg.Source: APPENDIX · p. 17
  • Equation (A.1) expresses controller-frame dq variables as a phase-offset transformation of system-frame dq variables.The relation is x^c_dq = e^-jδ x^s_dq.Source: APPENDIX · p. 17
  • During grid disturbances, Eref saturation when Vomag differs from Vref naturally bypasses the RPC loop and SLVM control.The appendix refers readers to for details on the power-plant and adaptive virtual-impedance modeling.Source: APPENDIX · p. 17, 1) Small-signal modelling of the power… · p. 17
  • The appendix models power-plant and adaptive virtual-impedance dynamics, Voq feedforward, Vomag-based Pref adjustment, and Pref-Iomag droop before combining them into the complete fast-IVS small-signal model.The complete model combines equations (A.4), (A.7), (A.12), and (A.17).Source: 2) Small-signal modelling of Voq… · pp. 17–18, 5) Complete small-signal model · p. 18
Loading 2212.03053v2…