With the rapid integration of photovoltaic (PV) systems into power grids worldwide, the dynamic behavior of grid-connected solar inverters during grid disturbances has become a critical research focus. Large-scale PV plants, characterized by both decentralized low-voltage integration and centralized medium- to high-voltage connections, significantly impact grid stability. Consequently, grid codes in many countries now mandate low-voltage ride-through (LVRT) capabilities for solar inverters, requiring them to remain connected and support the grid during voltage dips. This article delves into the dynamic characteristics of grid-connected solar inverters under symmetrical and asymmetrical grid faults, providing a comprehensive theoretical and simulation-based analysis. Understanding these dynamics is essential for designing robust LVRT control strategies and ensuring reliable operation of PV power plants.
Solar inverters serve as the interface between PV arrays and the grid, converting DC power from solar panels into AC power synchronized with the grid. Unlike wind turbines, PV systems lack rotational inertia, making their fault responses primarily dependent on the inverter’s power electronic controls. The key challenge during grid faults is managing the DC-link voltage and AC output current within safe limits, as excessive currents can trigger protection mechanisms and lead to disconnection. This analysis focuses on single-stage, grid-connected solar inverters, which are prevalent in large-scale applications due to their simplicity and cost-effectiveness. We derive mathematical models in positive and negative synchronous rotating reference frames, examine power flow under fault conditions, and highlight the implications for maximum power point tracking (MPPT) control. The insights gained can guide the development of advanced fault-tolerant strategies for modern solar inverters.
Control Strategies for Grid-Connected Solar Inverters
The typical topology of a single-stage grid-connected solar inverter consists of a PV array, a DC-link capacitor, a voltage-source inverter (VSI), and a step-up transformer for isolation and voltage matching. The PV array charges the DC-link capacitor, and the VSI converts DC to AC power using pulse-width modulation (PWM). A key component is the anti-reverse diode, which prevents back-feeding from the grid to the PV array. The control strategy commonly employed is based on grid voltage-oriented vector control in the synchronous rotating dq-frame. This approach decouples active and reactive power control, facilitating independent regulation of current components.
The control block diagram includes an outer MPPT loop that adjusts the DC-link voltage reference (Vdc,ref) based on measurements of DC voltage and current. The error between Vdc,ref and the actual DC-link voltage (Vdc) passes through a PI controller to generate the d-axis current reference (id,ref), which governs active power output. The q-axis current reference (iq,ref) is typically set to zero for unity power factor but can be adjusted for reactive power support during faults. The inner current loop uses PI controllers to compute the inverter output voltage components (vd and vq) based on the errors between reference and actual currents (id and iq). These voltage references are then transformed to the stationary frame for PWM generation. To maintain output voltage consistency despite varying DC-link voltages, normalization is applied in SPWM or accounted for in SVPWM. Current limiting is crucial; during faults, id,ref is constrained between 0 and 1.1 per unit (p.u.) to prevent overcurrent, ensuring the solar inverter remains within its thermal limits.
The mathematical model in the dq-frame is essential for analyzing solar inverter dynamics. The grid voltage and inverter output current are transformed using the Park transformation. Under balanced conditions, the grid voltage in the dq-frame is:
$$e_d = \sqrt{2}E, \quad e_q = 0$$
where E is the RMS phase voltage. The active power output of the solar inverter is:
$$p = \frac{3}{2} e_d i_d$$
This formulation simplifies control design but requires adaptation under unbalanced faults, where negative-sequence components emerge. The behavior of solar inverters under such conditions is explored in subsequent sections.
Dynamic Characteristics Under Symmetrical Grid Faults
Symmetrical faults, such as three-phase short circuits, cause equal voltage dips in all phases without phase angle jumps. Assume the grid voltage drops to a fraction of its pre-fault value, with Ef representing the post-fault RMS voltage. In the dq-frame, this translates to:
$$e_{fd} = \sqrt{2}E_f, \quad e_{fq} = 0$$
The maximum active power deliverable by the solar inverter during the fault is limited by the current constraint:
$$p_{f,max} = 1.5 e_{fd} i_{d,max}$$
where id,max = 1.1 p.u. The dynamic response depends on the pre-fault operating point relative to pf,max. Under high solar irradiance, the PV array generates power exceeding pf,max. When the fault occurs, the solar inverter cannot inject all available power into the grid, causing a power imbalance. The DC-link voltage rises as the capacitor charges, and id increases to id,max due to current limiting. MPPT control is suspended, and the system settles at a new operating point where p’ = pf,max. After fault clearance, the solar inverter resumes MPPT and returns to its original state.
Conversely, under low irradiance, the pre-fault power is below pf,max. The fault does not saturate the current limit, so id increases but remains within bounds. MPPT control continues uninterrupted, and the DC-link voltage remains stable. This highlights how the power output capability of solar inverters dictates PV array operation during symmetrical faults. Table 1 summarizes these scenarios.
| Condition | Pre-fault Power vs. pf,max | DC-link Voltage Response | MPPT Status | Post-fault Recovery |
|---|---|---|---|---|
| High Irradiance | p > pf,max | Increases | Suspended | Returns to MPPT |
| Low Irradiance | p < pf,max | Stable | Active | Unaffected |
The mathematical analysis can be extended using differential equations. The DC-link dynamics are governed by:
$$C \frac{dV_{dc}}{dt} = I_{pv} – I_{dc}$$
where C is the DC-link capacitance, Ipv is the PV array current, and Idc is the inverter input current. During faults, Idc is constrained, leading to voltage fluctuations. The response time depends on C and the power mismatch. Larger capacitors can mitigate voltage swings but increase cost and size, presenting a design trade-off for solar inverters.
Dynamic Characteristics Under Asymmetrical Grid Faults
Asymmetrical faults, such as single-phase or phase-to-phase faults, result in unbalanced grid voltages with positive and negative-sequence components. Since solar inverters are typically three-wire systems, zero-sequence currents are absent. The voltage and current vectors can be decomposed into positive and negative sequences, analyzed in positive (dq+) and negative (dq–) synchronous rotating frames. Let φ represent the angle between the negative-sequence d-axis and positive-sequence d-axis, which varies at twice the grid frequency (2ω1). The grid voltage and inverter current in the dq+ frame contain DC components from positive sequences and AC components at 2ω1 from negative sequences.
The instantaneous active and reactive power outputs of the solar inverter under unbalanced conditions are derived using the instantaneous power theory. Expressing voltages and currents in sequence components:
$$e_d = e_d^+ + e_d^- \cos(2\omega_1 t) – e_q^- \sin(2\omega_1 t)$$
$$e_q = e_q^+ + e_d^- \sin(2\omega_1 t) + e_q^- \cos(2\omega_1 t)$$
$$i_d = i_d^+ + i_d^- \cos(2\omega_1 t) – i_q^- \sin(2\omega_1 t)$$
$$i_q = i_q^+ + i_d^- \sin(2\omega_1 t) + i_q^- \cos(2\omega_1 t)$$
The active power p(t) and reactive power q(t) are:
$$p(t) = p_0 + p_{c2} \cos(2\omega_1 t) + p_{s2} \sin(2\omega_1 t)$$
$$q(t) = q_0 + q_{c2} \cos(2\omega_1 t) + q_{s2} \sin(2\omega_1 t)$$
where p0 and q0 are average powers, and pc2, ps2, qc2, qs2 are amplitudes of double-frequency oscillations. These coefficients are:
$$p_0 = 1.5(e_d^+ i_d^+ + e_q^+ i_q^+ + e_d^- i_d^- + e_q^- i_q^-)$$
$$p_{c2} = 1.5(e_d^+ i_d^- + e_q^+ i_q^- + e_d^- i_d^+ + e_q^- i_q^+)$$
$$p_{s2} = 1.5(e_d^+ i_q^- – e_q^+ i_d^- – e_d^- i_q^+ + e_q^- i_d^+)$$
$$q_0 = 1.5(e_q^+ i_d^+ – e_d^+ i_q^+ + e_q^- i_d^- – e_d^- i_q^-)$$
$$q_{c2} = 1.5(e_q^+ i_d^- – e_d^+ i_q^- + e_q^- i_d^+ – e_d^- i_q^+)$$
$$q_{s2} = 1.5(e_d^+ i_d^- + e_q^+ i_q^- – e_d^- i_d^+ – e_q^- i_q^+)$$
Neglecting inverter losses, the DC-side power equals the AC-side active power:
$$V_{dc} I_{dc} = p(t) = p_0 + p_{c2} \cos(2\omega_1 t) + p_{s2} \sin(2\omega_1 t)$$
This reveals that double-frequency ripples appear in both DC-link voltage and power. The PV array, acting as a non-ideal source, influences the distribution of ripple between voltage and current. Since solar inverters are voltage-source types, the voltage ripple is more critical, potentially affecting modulation and control.
Under conventional control strategies, only positive-sequence currents are regulated. The negative-sequence voltage at the point of interconnection (POI) drives large negative-sequence currents through the filter impedance, which are uncontrolled. While the dq-current components may stay within limits, individual phase currents can exceed rated values, risking overcurrent protection. For example, if id and iq are limited to 1.1 p.u., the phase currents in a-b-c coordinates might surpass this due to negative-sequence addition. This underscores a key limitation of conventional control in solar inverters during asymmetrical faults.
Moreover, the presence of double-frequency components in ed, eq, id, and iq challenges PI controllers, which cannot track AC signals without steady-state error. The computed vd and vq contain second harmonics, leading to third-harmonic voltages and currents after inverse transformation. Importantly, this harmonic distortion is not caused by DC-link voltage ripple but by the controller’s inherent limitation. Even if the DC-link voltage is steady, the PI controllers still produce erroneous outputs for AC quantities. To mitigate this, advanced strategies like resonant controllers or dual-sequence control are needed for solar inverters.
The impact on MPPT control is minimal. Perturb-and-observe MPPT algorithms operate on time scales much longer than the double-frequency period. Voltage and power averages over hundreds of milliseconds smooth out the ripples, ensuring MPPT decisions are unaffected. Thus, solar inverters can maintain maximum power extraction even during unbalanced faults, provided current limits are not breached.
Table 2 compares key aspects of symmetrical and asymmetrical fault responses for solar inverters.
| Aspect | Symmetrical Faults | Asymmetrical Faults | |||
|---|---|---|---|---|---|
| Voltage Components | Positive-sequence only | Positive and negative sequences | |||
| Current Control | Effective with dq-limits | Phase currents may exceed limits | |||
| Power Oscillations | None (steady DC power) | Double-frequency ripples in AC/DC power | DC-link Voltage | May rise if power-limited | Ripples at 2ω1 |
| Harmonic Generation | Negligible | Third harmonics due to PI limitations | |||
| MPPT Interference | Possible suspension | Minimal effect |
Simulation Verification and Case Studies
To validate the theoretical analysis, detailed simulations were conducted using a model of a 500 kW grid-connected solar inverter. The parameters are listed in Table 3, reflecting typical large-scale PV systems. The inverter employs the conventional vector control strategy with current limiting, as described earlier. Faults are applied at the POI, and dynamic responses are recorded for both symmetrical and asymmetrical scenarios.
| Parameter | Value |
|---|---|
| Rated Power | 500 kW |
| Grid Frequency | 50 Hz |
| DC-link Capacitance | 18,000 μF |
| AC Filter Inductance | 0.17 mH |
| AC Filter Capacitance | 200 μF |
| Transformer Ratio | 270 V / 400 V (DyN1) |
| Current Limit (id,max) | 1.1 p.u. |
For symmetrical faults, a three-phase fault at t = 4 s reduces the POI voltage to 0.5 p.u. for 2 seconds. Under high irradiance (1500 W/m²), the pre-fault power is 345 kW, exceeding the fault-time maximum of 275 kW. The DC-link voltage rises from 750 V to 780 V, and id saturates at 1.1 p.u., confirming power limitation. MPPT is halted during the fault. Under low irradiance (1000 W/m²), the pre-fault power is 235 kW, below the limit, so id increases to 0.9 p.u. without voltage surge, and MPPT continues. These results align with the theoretical predictions, demonstrating how solar inverters adapt based on available power.
For asymmetrical faults, a single-phase fault is simulated, resulting in positive-sequence voltage of 0.75 p.u. and negative-sequence voltage of 0.25 p.u. at the POI. The conventional control strategy leads to significant negative-sequence currents, causing double-frequency oscillations in DC-link voltage (ripple amplitude of ~10 V) and power. Phase currents show imbalance, with one phase exceeding 1.2 p.u., highlighting the inadequacy of dq-current limiting. Additionally, FFT analysis reveals third-harmonic distortion (about 5% THD) in the inverter output current, attributable to PI controller errors. When a modified control with a notch filter to remove second harmonics from vd and vq is applied, the third harmonics vanish, proving their origin from control rather than DC ripple. MPPT remains effective throughout, with the DC-link voltage trend undisturbed by the oscillations.
These simulations underscore the need for enhanced control schemes in solar inverters to handle unbalanced conditions. Strategies such as negative-sequence current injection, resonant controllers, or model predictive control can mitigate phase overcurrents and reduce harmonics, improving LVRT performance.
Conclusion
This analysis provides a thorough examination of the dynamic characteristics of grid-connected solar inverters under power grid faults. For symmetrical faults, the behavior of solar inverters is determined by their power output capability relative to pre-fault generation; high irradiance leads to DC-link voltage rise and MPPT suspension, while low irradiance allows uninterrupted operation. For asymmetrical faults, conventional control strategies prove inadequate, as negative-sequence currents cause phase overcurrents and double-frequency ripples in DC quantities. The resulting third-harmonic distortion stems from PI controllers’ inability to track AC signals, not from DC voltage fluctuations. Importantly, MPPT control is largely unaffected by these dynamics due to its slower sampling.
The findings emphasize that solar inverters must incorporate advanced control mechanisms to comply with stringent grid codes. Future work could explore dual-sequence control, active filtering, and adaptive current limiting to enhance fault ride-through capabilities. As PV penetration grows, understanding and optimizing the dynamic response of solar inverters will be crucial for grid stability and renewable energy integration. This research contributes to that goal by offering insights into fault-induced behaviors and guiding the development of robust solar inverter technologies.