In modern power systems, large-scale photovoltaic (PV) integration demands that grid-connected solar inverters maintain operation during grid faults. The low-voltage ride-through (LVRT) capability is essential to prevent widespread disconnection and support grid recovery. This paper presents a comprehensive LVRT control strategy tailored for a two-stage solar inverter topology composed of a front-end Boost converter and a rear-end DC/AC inverter. The proposed method focuses on rapid power balancing and coordinated active/reactive current injection without additional hardware. By modifying the Boost control mode during voltage sags and introducing a feedforward term based on the pre-fault maximum power point voltage, the solar inverter can quickly regulate the PV array output power. Simultaneously, the inverter current references are adjusted according to the voltage drop depth to provide reactive power support. Detailed mathematical derivations, control block descriptions, and simulation results based on a 100 kW two-stage solar inverter model are provided. Tables summarizing key parameters and performance comparisons are included to validate the effectiveness of the proposed approach.
Introduction
Grid-connected solar inverters are critical components in renewable energy systems. With increasing penetration levels, grid codes require solar inverters to remain connected during voltage dips and even contribute to voltage support. Two-stage solar inverters, consisting of a Boost stage for maximum power point tracking (MPPT) and an inverter stage for grid interface, face particular challenges during low-voltage faults. The power imbalance between the DC side and AC side can cause DC-link overvoltage and inverter overcurrent, leading to disconnection. Conventional LVRT solutions often rely on additional circuits such as choppers or energy storage, which increase cost and complexity. This work proposes a control-level approach that achieves both power balance and reactive support without extra hardware. The core idea is to alter the Boost control objective from MPPT to DC-link voltage regulation during faults, using a feedforward of the pre-fault PV array voltage to rapidly reduce PV output power. Meanwhile, the inverter control switches from voltage-loop-based active current generation to a fixed current reference scheme that distributes current between active and reactive components according to the sag depth. The effectiveness is demonstrated through simulation of a 100 kW two-stage solar inverter system.
System Topology and Conventional Control
The two-stage solar inverter under study is illustrated conceptually. The front stage is a Boost converter that elevates the PV array voltage to a stable DC-link voltage. The rear stage is a three-phase voltage source inverter connected to the grid via an LCL filter. The DC-link capacitor provides energy buffering. Under normal conditions, the Boost stage operates with MPPT using a perturbation and observation (P&O) method, maintaining the PV array at its maximum power point. The inverter stage controls the DC-link voltage through an outer voltage loop and the grid currents through inner current loops in the synchronous reference frame.
The conventional Boost control structure is given by the following relationship:
$$
u_{pv,ref} = u_{pv}^{*}
$$
where \(u_{pv}^{*}\) is obtained from MPPT algorithm. The inverter current control employs PI regulators:
$$
u_d = \left(k_{ip} + \frac{k_{ii}}{s}\right)(i_d^{*} – i_d) – \omega L i_q + e_d
$$
$$
u_q = \left(k_{ip} + \frac{k_{ii}}{s}\right)(i_q^{*} – i_q) + \omega L i_d + e_q
$$
Here, \(i_d^{*}\) is generated from the DC-link voltage PI controller, and \(i_q^{*}\) is normally set to zero for unity power factor. The equations use standard notation: \(k_{ip}, k_{ii}\) are PI gains, \(\omega\) is grid frequency, \(L\) is the filter inductance (neglecting capacitor branch), and \(e_d, e_q\) are grid voltage components.
| Parameter | Symbol | Value |
|---|---|---|
| Rated power | \(P_N\) | 100 kW |
| DC-link voltage reference | \(u_{dc,ref}\) | 800 V |
| DC-link capacitance | \(C\) | 6000 \(\mu\)F |
| Grid nominal voltage (L-N rms) | \(u_N\) | 220 V |
| Grid frequency | \(f\) | 50 Hz |
| Inverter rated current | \(I_N\) | 200 A |
| LCL filter: inverter-side inductance | \(L_1\) | 0.4 mH |
| LCL filter: grid-side inductance | \(L_2\) | 0.05 mH |
| LCL filter: filter capacitance | \(C_f\) | 10 \(\mu\)F |
| Damping resistor | \(R_d\) | 1 \(\Omega\) |
Impact of Low-Voltage Faults
When a balanced voltage sag occurs at the point of common coupling (PCC), the injected AC power \(P_{out}\) suddenly decreases due to both voltage reduction and current limitation. The PV array, still operating at MPPT, continues to supply maximum power \(P_{pv}\). The power imbalance \(\Delta P\) charges the DC-link capacitor, causing voltage rise:
$$
\Delta P = P_{pv} – P_{out} = \frac{1}{2}C\frac{u_{dc}'(t)^2 – u_{dc}(0)^2}{\Delta t}
$$
where \(u_{dc}(0)\) is the pre-fault DC voltage and \(u_{dc}'(t)\) is the instantaneous voltage during fault. Without intervention, the DC-link voltage will exceed safe limits. Moreover, the inverter current controller will increase \(i_d^{*}\) to regulate the DC voltage, potentially causing overcurrent. Hence, two key issues must be addressed: (1) rapidly reducing PV output power to match the reduced AC power, and (2) limiting inverter current within safe bounds while providing reactive support.
Proposed LVRT Control Strategy
Boost Control Modification
During voltage sags, the conventional MPPT algorithm is abandoned and replaced by a new control loop that regulates the PV array voltage to a value higher than the maximum power point voltage \(u_{max}\). This is achieved by introducing a DC-link voltage controller whose output is added to the stored pre-fault \(u_{max}\). The modified Boost reference becomes:
$$
u_{pv,ref} = u_{max} + \Delta u
$$
$$
\Delta u = k_{p,dc} (u_{dc} – u_{dc,ref}) + k_{i,dc} \int (u_{dc} – u_{dc,ref}) dt
$$
When the DC-link voltage rises due to sag, the PI controller outputs a positive \(\Delta u\), shifting the PV operating point to the right of the MPP, thereby reducing PV output power. This approach is preferred over moving left because the right-side power-voltage slope is steeper, allowing faster power reduction. Moreover, moving right reduces the Boost duty cycle, simplifying control. The pre-fault \(u_{max}\) is stored during normal MPPT operation and held constant during the fault.
Inverter Current Reference Generation
During LVRT, the inverter must limit its total current magnitude to \(I_{max}\) (typically 1.1 × \(I_N\)) while injecting reactive current proportional to the voltage sag depth. Let \(k = u_f / u_N\) be the residual voltage ratio. The active current reference \(i_d^{*}\) is linearly reduced from its pre-fault value to a value \(k I_{max}\) after a brief hold period (3 ms) and a linear ramp down. The reactive current reference \(i_q^{*}\) is then computed as:
$$
i_q^{*} = \sqrt{I_{max}^2 – (i_d^{*})^2}
$$
The active current reduction profile is expressed as:
$$
i_d^{*}(t) =
\begin{cases}
i_{d,pre} & t_0 \leq t < t_0 + \Delta t_{hold} \\
i_{d,pre} – r (t – t_0 – \Delta t_{hold}) & t_0 + \Delta t_{hold} \leq t < t_1 \\
k I_{max} & t \geq t_1
\end{cases}
$$
where \(r\) is the reference decay rate, \(\Delta t_{hold} = 3\) ms, and \(t_1\) is the moment when \(i_d^{*}\) reaches \(k I_{max}\). During fault recovery, the references are smoothly returned to normal operation using a similar ramp.
The overall current control during LVRT is summarized by the following block description: the outer DC-link voltage loop is disabled; instead, fixed references \(i_d^{*}\) and \(i_q^{*}\) are directly supplied to the inner current regulators. This decouples the inverter from DC voltage dynamics, preventing overcurrent while enabling reactive support.
Simulation Verification
A 100 kW two-stage solar inverter model is built in PSCAD/EMTDC using the parameters listed in Table 1. The grid voltage is set to 1 p.u. and a balanced three-phase sag to 0.2 p.u. occurs at 0.3 s, lasting for 0.1 s. The PV array characteristics are assumed constant during the fault (irradiance steady). The proposed LVRT strategy is activated when PCC voltage drops below 0.9 p.u.
Figure 1 below shows a typical solar inverter unit used in such simulations.
Results and Discussion
The simulation results demonstrate the effectiveness of the proposed control. Key performance metrics are summarized in Table 2.
| Parameter | Without LVRT control | With proposed LVRT control |
|---|---|---|
| Maximum DC-link voltage (V) | >950 (overvoltage trip) | 861.5 |
| Peak inverter current (A) | >250 (overcurrent) | 215 (limited) |
| PV output power after transient (kW) | 100 (unchanged) | ~20 (reduced) |
| PCC voltage during fault (V rms) | 41.91 | 46.88 |
| Reactive current injection (A) | 0 | ~180 |
| Voltage recovery time (ms) | N/A (disconnected) | 20 |
The DC-link voltage peaked at 861.5 V, which is within the safe operating range of typical 800 V systems (usually rated up to 900 V). The PV output power quickly dropped from 100 kW to about 20 kW within 6 ms, matching the reduced AC power capability. The inverter current was limited to 215 A (slightly above rated but below the 1.1 p.u. limit). The reactive current injection reached approximately 180 A, raising the PCC voltage from 41.91 V to 46.88 V, providing clear voltage support. Without the proposed control, the conventional strategy would have caused DC overvoltage and inverter overcurrent, leading to protective disconnection.
Additional simulations were performed for different sag depths (0.5 p.u. and 0.3 p.u.). The results are listed in Table 3.
| Sag depth (p.u.) | Peak DC-link voltage (V) | Reactive current (A) | PCC voltage with support (V rms) | PCC voltage without support (V rms) |
|---|---|---|---|---|
| 0.5 | 825 | 100 | 115 | 110 |
| 0.3 | 840 | 150 | 72 | 66 |
| 0.2 | 861.5 | 180 | 46.88 | 41.91 |
These tables confirm that the proposed strategy consistently maintains DC-link voltage within safe limits while providing reactive support proportional to sag severity. The voltage support effect is more pronounced for deeper sags due to higher reactive current injection.
Conclusion
This paper has presented a control strategy for low-voltage ride-through of a two-stage solar inverter that coordinates active power reduction and reactive power injection. By modifying the Boost control to shift the PV operating point to the right of MPPT using a feedforward of the pre-fault voltage, the solar inverter quickly balances power between the PV array and the grid, preventing DC-link overvoltage. Simultaneously, the inverter current references are set based on voltage sag depth, ensuring the solar inverter provides reactive support while staying within current limits. Simulation results on a 100 kW system demonstrate that the strategy effectively limits DC-link voltage to safe values, avoids overcurrent, and improves PCC voltage during faults. No additional hardware is required, making the approach cost-effective and easily implementable in existing solar inverter controllers.