With the rapid increase in the proportion of renewable energy in power systems, the overall inertia and frequency regulation capabilities of the grid have significantly declined. In the current context of large-scale renewable energy bases and high-voltage direct current transmission, grid faults can easily trigger widespread disconnections of renewable sources, leading to stability issues. To address this, grid-forming solar inverters have emerged as a promising solution, enabling active support for the grid and independent network construction under certain conditions. This article delves into the control strategy and implementation of grid-forming solar inverters without energy storage support, based on power self-synchronization principles. Through systematic testing, including active grid support, grid-connected to islanding transition, black start, and multi-unit parallel operation, the effectiveness of this approach is demonstrated. The results highlight the advantages of solar inverters in enhancing grid friendliness and providing independent voltage and frequency support, offering a technical pathway for large-scale renewable energy integration.
The core of grid-forming solar inverters lies in their ability to mimic the behavior of traditional synchronous generators, providing inherent grid-friendly characteristics. Unlike conventional grid-following solar inverters, which rely on phase-locked loops (PLLs) for synchronization, grid-forming solar inverters utilize power self-synchronization to autonomously establish voltage and frequency. This eliminates the dependency on external grid signals, enhancing stability in weak grid conditions. The control strategy for solar inverters involves reserving spare capacity to support active and reactive power, allowing them to participate in grid regulation and operate independently during outages. Key aspects include power synchronization, virtual impedance, and current control loops, which collectively enable solar inverters to emulate synchronous machine dynamics.
The topological structure of a grid-forming solar inverter typically includes a photovoltaic (PV) equivalent DC source, a three-phase inverter bridge, and an output filter. The DC source represents the PV array, while the inverter and filter simulate the internal impedance and terminal voltage of a synchronous generator. In grid-connected mode, the inverter’s output voltage aligns with the grid voltage, whereas in islanded mode, it independently controls the local voltage and frequency. The control system for solar inverters is designed to achieve frequency-active power and voltage-reactive power droop characteristics, similar to synchronous generators. The power control loop equations are fundamental to this approach, as shown below:
$$ P_{ref} – K_i \int (\omega – \omega_{ref}) dt – P_e = J \omega \frac{d\omega}{dt} $$
$$ E_m = \frac{D_q (U_n – U_0) + Q_{ref} – Q_e}{K s} $$
Here, $P_{ref}$ is the reference mechanical power (equivalent to the PV output at rated grid frequency $\omega_{ref}$), $Q_{ref}$ is the reactive power reference (corresponding to rated voltage $U_n$), $D_q$ is the voltage-reactive power droop coefficient, $P_e$ and $Q_e$ are the output active and reactive power of the solar inverter, $J$ is the virtual inertia, $K_d$ is the mechanical damping coefficient, and $K$ is an integral constant. The Laplace operator $s$ denotes the derivative in the frequency domain. For solar inverters, the PV array’s output power serves as $P_{ref}$, and a DC voltage control loop is incorporated to maintain stability, given by:
$$ P_{ref} = (U_{dc\_ref} – U_{dc}) G_{dc}(s) $$
where $U_{dc\_ref}$ is the DC bus voltage reference from the maximum power point tracking (MPPT) algorithm, and $G_{dc}(s)$ is a proportional-integral (PI) controller. This ensures that the solar inverter operates at the maximum power point while reserving capacity for grid support. The spare power capacity $\Delta P_e$ is defined as:
$$ \Delta P_e = m P_{max} $$
with $m$ being the spare coefficient (0 ≤ m ≤ 1). This allows the solar inverter to release reserved power during frequency deviations, participating in primary frequency regulation. The reactive power output is constrained by the inverter’s rated capacity $S_N$ and active power output, satisfying $S_N^2 \geq P_e^2 + Q_e^2$. Under steady-state conditions, the effective boundary for reactive power is approximated as:
$$ Q_e \leq \sqrt{S_N^2 – [(1-m) P_{max} + D_p (\omega_{ref} – \omega_g)]^2} $$
where $D_p$ is the frequency-active power droop coefficient. To achieve power self-synchronization, the solar inverter’s control includes virtual impedance and current inner loops. The equations for the current reference in the dq-frame are:
$$ L_v \frac{di_{Ld}^*}{dt} = e_d – u_{gd} – R_v i_{Ld} + \omega L_v i_{Lq} $$
$$ L_v \frac{di_{Lq}^*}{dt} = e_q – u_{gq} – R_v i_{Lq} – \omega L_v i_{Ld} $$
where $L_v$ and $R_v$ are the virtual inductance and resistance, $i_{Ld}$ and $i_{Lq}$ are the dq-axis inductor currents, $u_{gd}$ and $u_{gq}$ are the grid voltage components, and $e_d$ and $e_q$ are the internal voltage components derived from the power loop. The current control loop outputs modulation signals $u_{d}$ and $u_{q}$ as:
$$ u_d = K_{pc} (i_{Ld}^* – i_{Ld}) + K_{ic} \int (i_{Ld}^* – i_{Ld}) dt + k u_{gd} – \omega L_f i_{Lq} $$
$$ u_q = K_{pc} (i_{Lq}^* – i_{Lq}) + K_{ic} \int (i_{Lq}^* – i_{Lq}) dt + k u_{gq} + \omega L_f i_{Ld} $$
with $K_{pc}$ and $K_{ic}$ being the PI gains, and $k$ the voltage feedforward coefficient. This control structure enables solar inverters to autonomously synchronize with the grid without PLLs, improving dynamic response. The overall control block diagram for grid-forming solar inverters integrates these elements, emphasizing power synchronization and virtual impedance.
In practical engineering, the implementation of grid-forming solar inverters was tested in a renewable energy base. The setup involved multiple solar inverters connected to a 35 kV grid, with a 5 MW energy storage system emulating grid fluctuations. Key parameters for the solar inverters are summarized in the table below:
| Parameter | Value |
|---|---|
| Rated Power | 500 kW |
| DC Voltage Range | 450-820 V |
| Filter Inductance $L_f$ | 0.15 mH |
| Filter Capacitance $C_f$ | 600 μF |
| Active Power Droop Coefficient $D_p$ | 18,000 |
| Reactive Power Droop Coefficient $D_q$ | 20,000 |
| Virtual Inertia $J$ | 0.133 kg·m² |
Tests on active grid support demonstrated the solar inverter’s ability to respond to frequency and voltage variations. For instance, when the grid frequency dropped by 0.5 Hz, the solar inverter rapidly released reserved active power of 40 kW, with a response time under 200 ms. Similarly, for a voltage dip of 0.05 pu, the solar inverter increased reactive power output by 93 kvar. The table below quantifies the performance of solar inverters in these tests:
| Test Scenario | Response Time | Power Adjustment |
|---|---|---|
| Frequency Drop (0.5 Hz) | < 200 ms | +40 kW Active Power |
| Frequency Rise (0.5 Hz) | < 200 ms | -40 kW Active Power |
| Voltage Dip (0.05 pu) | ≤ 200 ms | +93 kvar Reactive Power |
| Voltage Rise (0.05 pu) | ≤ 200 ms | -93 kvar Reactive Power |
Transition tests from grid-connected to islanded mode showed that solar inverters could maintain stable voltage and frequency for local loads. During the switch, the voltage fluctuation was limited to 0.05 pu, and multiple solar inverters operated in parallel with seamless transition. Black start tests confirmed that a single solar inverter could independently establish a 35 kV grid voltage and frequency, with total harmonic distortion below acceptable limits. The process involved soft starting, with voltage buildup achieved within 200 ms. Multi-unit parallel operation further validated the autonomous current sharing capability of solar inverters, with power sharing accuracy exceeding 90%. The dynamic and steady-state performance of solar inverters in parallel operation is summarized below:
| Parameter | Value |
|---|---|
| Number of Units | 5 |
| Voltage Stability | ±0.05 pu |
| Frequency Stability | ±0.2 Hz |
| Current Sharing Accuracy | > 90% |
The control strategy for solar inverters also involves periodic switching between MPPT and power reserve modes to adapt to environmental changes. The power reference $P_{ref}$ is adjusted based on the spare coefficient $m$, as given by:
$$ P_{ref} = (1 – m) P_{max} $$
This ensures that solar inverters maintain sufficient reserve for grid support while maximizing energy harvest. The synchronization mechanism for solar inverters relies on the power-angle relationship, derived from the swing equation:
$$ \frac{d\delta}{dt} = \omega – \omega_g $$
where $\delta$ is the power angle, and $\omega_g$ is the grid frequency. By integrating this with virtual impedance, solar inverters achieve self-synchronization without external synchronization signals. The stability of solar inverters in weak grids is enhanced by the damping provided by virtual impedance, which can be expressed as:
$$ Z_v = R_v + j \omega L_v $$
In conclusion, grid-forming solar inverters based on power self-synchronization offer a robust solution for renewable energy integration. The control strategy enables solar inverters to provide active grid support, independent operation, and black start capabilities without energy storage. Experimental results confirm the superior performance of solar inverters in terms of response time, voltage and frequency stability, and grid friendliness. Future work could focus on optimizing the control parameters for larger-scale deployments and integrating advanced forecasting for spare capacity management. Overall, solar inverters play a critical role in the transition to sustainable power systems, and their continued development will be essential for achieving high renewable energy penetration.