In modern power systems, the integration of renewable energy sources, particularly solar power, has become increasingly prevalent. Solar inverters play a critical role in converting DC power from photovoltaic arrays into AC power suitable for grid injection. However, the intermittent nature of solar energy and the grid’s stability requirements pose significant challenges. One such challenge is the ability of solar inverters to remain connected and supportive during grid voltage dips, a capability known as Low Voltage Ride-Through (LVRT). As grid codes worldwide mandate LVRT compliance, developing effective control strategies for solar inverters has become paramount. In this article, I propose a novel voltage direct feed-forward control method to enhance LVRT performance in solar inverters, addressing issues such as overcurrent during voltage sags and providing reactive power support. This approach leverages real-time grid voltage measurements to improve dynamic response and ensure safe operation under various fault conditions.
The importance of LVRT for solar inverters cannot be overstated. When grid voltage drops due to faults, if multiple solar inverters disconnect simultaneously, it can exacerbate grid instability, leading to potential blackouts. Therefore, solar inverters must not only stay connected during voltage dips but also contribute to grid recovery by supplying reactive power. Conventional control systems for solar inverters often rely on synchronous rotating coordinate transformations and proportional-integral (PI) regulators. However, these methods may suffer from delays and inadequate handling of unbalanced voltage sags, resulting in overcurrent and tripping. I will explore these limitations and introduce a solution based on direct voltage feed-forward, which minimizes response lag and enhances robustness.
To set the foundation, let’s consider the typical configuration of a grid-connected solar inverter system. A solar inverter consists of a DC-AC converter, output filters (inductors and capacitors), and control circuitry. The control system typically uses a Phase-Locked Loop (PLL) to synchronize with the grid voltage vector, enabling the decomposition of currents into active (d-axis) and reactive (q-axis) components for independent control. The standard control block diagram, as shown in conventional setups, includes MPPT for DC voltage regulation and current loops for power control. However, this system is primarily designed for balanced grid conditions and may falter during asymmetrical voltage dips, where negative-sequence components arise.
In unbalanced grid conditions, such as single-phase or two-phase voltage sags, the grid voltage contains both positive- and negative-sequence components. For a three-wire system, ignoring zero-sequence components, the unbalanced three-phase voltages can be expressed as:
$$V_a(t) = V_m^+ \sin(\omega t + \phi^+) + V_m^- \sin(-\omega t + \phi^-)$$
$$V_b(t) = V_m^+ \sin\left(\omega t + \phi^+ – \frac{2\pi}{3}\right) + V_m^- \sin\left(-\omega t + \phi^- – \frac{2\pi}{3}\right)$$
$$V_c(t) = V_m^+ \sin\left(\omega t + \phi^+ + \frac{2\pi}{3}\right) + V_m^- \sin\left(-\omega t + \phi^- + \frac{2\pi}{3}\right)$$
where $\omega$ is the grid angular frequency, $V_m^+$ and $V_m^-$ are the positive- and negative-sequence voltage amplitudes, and $\phi^+$ and $\phi^-$ are their initial phases. Through coordinate transformation, the voltage vector in the rotating frame becomes:
$$\mathbf{V}(t) = e^{j\omega t} \mathbf{V}_{dq}^+ + e^{-j\omega t} \mathbf{V}_{dq}^-$$
with $\mathbf{V}_{dq}^+ = V_d^+ + jV_q^+$ and $\mathbf{V}_{dq}^- = V_d^- + jV_q^-$. The presence of negative-sequence components introduces double-frequency oscillations in the d-q axes, complicating control and necessitating filtering, which can delay response.
To address this, dual synchronous rotating coordinate control has been proposed. This method employs separate positive- and negative-sequence frames to independently regulate currents, with feed-forward decoupling for each sequence. While effective in steady-state under unbalanced conditions, it still suffers from delays due to sampling and computation. The filtering of double-frequency components further exacerbates lag, leading to potential overcurrent during transient voltage sags. This highlights the need for a more responsive approach in solar inverters.
The overcurrent phenomenon during voltage dips is primarily attributed to the lag in grid voltage feed-forward. In a solar inverter, the output current is proportional to the voltage across the output inductor, given by:
$$i = \frac{1}{L} \int u \, dt$$
where $L$ is the inductance and $u$ is the voltage across it. If the inverter’s output voltage reference does not adjust quickly enough to match the sagging grid voltage, the voltage difference across the inductor spikes, causing excessive current. This initial surge can trigger protection mechanisms and disconnect the solar inverter, violating LVRT requirements. Therefore, reducing feed-forward delay is crucial for LVRT compliance in solar inverters.
I propose a voltage direct feed-forward control strategy that mitigates this delay. The strategy consists of two complementary approaches: Strategy 1 uses compensated instantaneous grid voltage values for feed-forward in steady-state, improving dynamic response without filtering double-frequency harmonics. Strategy 2 employs raw instantaneous grid voltage samples during fault detection to achieve minimal lag in the initial moments of a voltage dip. This combination ensures that the solar inverter adapts swiftly to grid changes, suppressing overcurrent and maintaining connectivity.
The hardware implementation involves a digital signal processor (DSP) and a field-programmable gate array (FPGA) for control. Grid voltage is sampled at high speed for both average values (synchronized with PWM cycles) and instantaneous values (microsecond-level averages). The instantaneous samples are used for rapid fault detection and feed-forward during LVRT events. The control algorithm continuously monitors grid voltage amplitude and negative-sequence components to identify asymmetrical sags. Upon detection, the system switches to the direct feed-forward mode, adjusting the inverter’s output voltage reference in real time.
For current control, we implement dual-sequence regulation. The negative-sequence currents $I_d^-$ and $I_q^-$ are extracted via coordinate transformation and regulated to zero using PI controllers to minimize grid harmonics. The outputs $\Delta U_d^-$ and $\Delta U_q^-$ are transformed back to three-phase adjustments $\Delta U_{abc}^{-*}$. Similarly, positive-sequence current regulators produce $\Delta U_{abc}^{+*}$. These are combined to form the total voltage reference adjustment $\Delta U_{abc}^{*}$. Additionally, during LVRT, the active current reference is limited to 80% of rated value, and the reactive current reference is set to maximize reactive power support, calculated as:
$$I_q^* = \sqrt{(I_{\text{rated}})^2 – (I_d^*)^2}$$
where $I_{\text{rated}}$ is the rated current of the solar inverter. This ensures the solar inverter provides substantial reactive power to aid grid recovery while staying within thermal limits. To prevent overvoltage during grid recovery, the reactive support is phased out when grid voltage approaches 90% of nominal, avoiding post-fault surges.
The effectiveness of this control strategy is demonstrated through experimental results on a 250 kW solar inverter system. Tests were conducted using a grid simulator to emulate single-phase and two-phase voltage sags. The waveforms show that the solar inverter maintains stable output currents without overcurrent during dips, validating the LVRT capability. For instance, during a single-phase sag, the currents quickly stabilize at rated levels after a brief adjustment period. This performance underscores the robustness of the direct feed-forward approach for solar inverters in real-world scenarios.
To further illustrate the control concepts, let’s summarize key equations and comparisons in tables. Table 1 outlines the differences between conventional and proposed control methods for solar inverters during LVRT.
| Aspect | Conventional Control | Proposed Direct Feed-Forward Control |
|---|---|---|
| Feed-forward Source | Filtered grid voltage in d-q frame | Instantaneous grid voltage samples |
| Response Time | Delayed due to filtering and computation | Minimized (microsecond to half-PWM cycle lag) |
| Handling of Unbalanced Sags | Requires dual-sequence control with filtering | Uses direct samples, reducing harmonic issues |
| Overcurrent Suppression | Limited, especially during transients | Effective, with rapid voltage reference adjustment |
| Reactive Power Support | Often slow or inadequate | Maximized and phased out intelligently |
Table 2 lists common grid voltage dip types and their impact on solar inverter operation, emphasizing the need for advanced LVRT strategies.
| Dip Type | Description | Challenge for Solar Inverters |
|---|---|---|
| Single-Phase | Voltage drop in one phase | Unbalanced currents, negative-sequence components |
| Two-Phase | Voltage drop in two phases | Severe unbalance, high overcurrent risk |
| Three-Phase Symmetrical | Balanced drop in all phases | Less common, but can cause power loss |
| Three-Phase Asymmetrical | Unbalanced drop in all phases | Complex control due to mixed sequences |
The mathematical formulation of the control strategy can be extended. For the voltage direct feed-forward, we define the inverter output voltage reference as:
$$U_{abc}^* = U_{abc,\text{grid}} + \Delta U_{abc}^*$$
where $U_{abc,\text{grid}}$ is the instantaneous grid voltage vector sampled directly, and $\Delta U_{abc}^*$ is the adjustment from current regulators. This bypasses the need for sequence decomposition in the feed-forward path, reducing lag. The current regulators operate in the rotating frames, with equations:
$$\Delta U_d^+ = K_p (I_d^* – I_d^+) + K_i \int (I_d^* – I_d^+) dt – \omega L I_q^+$$
$$\Delta U_q^+ = K_p (I_q^* – I_q^+) + K_i \int (I_q^* – I_q^+) dt + \omega L I_d^+$$
for the positive sequence, and similarly for the negative sequence. The inclusion of cross-coupling terms ($\omega L I_q^+$ and $\omega L I_d^+$) enhances decoupling, but the direct feed-forward simplifies this by providing a more accurate grid voltage baseline.
In terms of implementation, the DSP executes control algorithms, while the FPGA handles logic and PWM generation. The sampling rate for instantaneous grid voltage is critical; we use a rate of several megahertz to capture rapid changes. This high-speed data enables the solar inverter to detect voltage dips within microseconds. Upon detection, the system immediately switches the feed-forward source to raw samples, ensuring the voltage reference aligns with the grid within a half-PWM cycle at worst. This rapid response is key to preventing overcurrent in solar inverters.
Furthermore, the control strategy includes adaptive current limiting. During normal operation, the solar inverter operates at maximum power point tracking (MPPT) or grid-dispatch modes. During LVRT, the active current reference $I_d^*$ is capped to prevent overloading, while the reactive current reference $I_q^*$ is increased to support grid voltage. The relationship is governed by:
$$I_d^{*2} + I_q^{*2} \leq I_{\text{rated}}^2$$
This ensures the solar inverter remains within its current rating while providing essential services. The transition back to normal mode is smoothed by gradually reducing reactive support as grid voltage recovers, preventing overvoltage issues.
Experimental validation involved testing the solar inverter under various dip scenarios. The results, as mentioned, show no overcurrent events and stable operation. For example, in a two-phase dip, the currents remained balanced after a brief transient, demonstrating the effectiveness of the negative-sequence control and direct feed-forward. These findings highlight that solar inverters equipped with this strategy can reliably meet stringent LVRT grid codes, enhancing grid resilience.
In conclusion, the voltage direct feed-forward control strategy offers a significant advancement for LVRT in solar inverters. By minimizing feed-forward delay and intelligently managing current references, it suppresses overcurrent during voltage dips and maximizes reactive power support. The use of high-speed sampling and dual-sequence control ensures robustness under unbalanced conditions. Experimental results confirm that solar inverters employing this method can safely ride through low voltage events, contributing to grid stability. As solar power penetration grows, such control strategies will be essential for reliable integration. Future work may explore integration with energy storage systems or multi-inverter coordination to further enhance performance.
To summarize the key points, I have presented a comprehensive approach to LVRT for solar inverters, emphasizing the role of direct voltage feed-forward. The strategy addresses common pitfalls in conventional systems and provides a practical solution validated through experiments. Solar inverters are pivotal in the renewable energy landscape, and advancing their grid-support capabilities is crucial for a sustainable power future. This work contributes to that goal by offering a control method that is both effective and implementable in real-world solar inverter applications.