The proliferation of distributed generation (DG), particularly from photovoltaic (PV) sources, is fundamentally reshaping modern power systems. As penetration levels rise, the behavior of these generating units during grid disturbances becomes critical for overall network stability. Grid operators worldwide have thus established stringent technical codes, mandating that solar inverters and other DG systems must remain connected and provide supportive functions during voltage sags or swells—a capability known as Fault Ride-Through (FRT).
My research focuses on the advanced control strategies required for three-phase, three-level solar inverters to comply with these demanding grid codes, specifically those modeled after the rigorous German Medium-Voltage Directive (MVD). This directive not only requires FRT but also stipulates active power management and dynamic reactive power support. The core challenge lies in maintaining stable operation of the solar inverter—preventing DC-link overvoltage and output current overshoot—while simultaneously injecting specified currents to support the grid during both symmetrical and asymmetrical faults. The typical two-stage topology for such a system comprises a front-end DC-DC boost converter for Maximum Power Point Tracking (MPPT) and a rear-end three-phase voltage source inverter for grid connection.
During a symmetrical grid fault (a balanced voltage sag), the primary control objectives shift. The immediate threat is a power imbalance: the PV array, via the MPPT controller, continues to deliver pre-fault power, while the inverter’s ability to inject power into the grid is diminished due to lower grid voltage. This surplus energy charges the DC-link capacitor, causing a rapid rise in DC-bus voltage that can trigger protection shutdowns. Furthermore, attempting to maintain pre-fault active power injection into a depressed voltage grid would lead to excessive current, exceeding the ratings of the semiconductor switches. Therefore, a coordinated control strategy encompassing both reactive power support and active power curtailment is essential.
The MVD specifies a characteristic for dynamic voltage support. Upon detecting a voltage dip beyond a 10% threshold, solar inverters must inject reactive current $(I_q)$ in proportion to the voltage deviation $(\Delta U)$. The required reactive current, as a percentage of rated current $(I_n)$, is given by:
$$ I_q^* = k \cdot \Delta U \cdot I_n $$
where $k \geq 2.0$ per unit. For instance, a 25% voltage dip requires at least 50% of the rated current as reactive current. This injection must commence within 20ms and be maintained for a specified duration after voltage recovery. This reactive current support aids in pulling the local grid voltage back towards its nominal value.
Concurrently, active power $(P)$ must be reduced to prevent overcurrent and stabilize the DC-link. The reference for the active current component $(I_d^*)$ is derived from a curtailed active power reference $(P_{ref}^*)$, which is managed by overriding the MPPT setpoint. The power balance equation governing the DC-link voltage $(e_d)$ is:
$$ C_{Bus} \frac{de_d}{dt} = P_{PV} – P_{out} \approx P_{MPPT} – \frac{3}{2}(u_d i_d + u_q i_q) $$
where $C_{Bus}$ is the DC-link capacitance, $P_{PV}$ is the input power from the PV array, and $P_{out}$ is the inverter output power expressed in the synchronous dq-reference frame. To prevent overcurrent, the total current magnitude must be limited to the rated value $I_{max}$:
$$ I_{out\_amp}^* = \sqrt{(I_d^*)^2 + (I_q^*)^2} \leq I_{max} $$
If the sum of the required $I_q^*$ and the initial $I_d^*$ exceeds this limit, the active power reference $P_{ref}^*$ (and consequently $I_d^*$) is reduced until the condition is met. The control structure thus integrates a DC-link voltage regulator that outputs a total current magnitude reference, which is then split into d and q components based on the grid voltage dip depth and the active power management scheme.
| Voltage Dip Magnitude $(\Delta U/U_n)$ | Required Reactive Current $(I_q^*/I_n)$ | Control Priority |
|---|---|---|
| $\Delta U < 10\%$ | 0 (or steady-state setpoint) | Steady-State Schedule |
| $10\% \leq \Delta U \leq 50\%$ | $k \cdot \Delta U$ $(k \geq 2.0)$ | Fault Mode (Priority) |
| $\Delta U > 50\%$ | $\geq 100\%$ | Fault Mode (Priority) |
The scenario becomes significantly more complex during asymmetrical faults, such as single-phase or phase-to-phase sags. Under such conditions, the grid voltages contain both positive and negative sequence components. If a standard controller designed for balanced conditions is used, it will generate highly distorted output currents rich in negative-sequence and second-harmonic components. This not only pollutes the grid but also introduces a destructive second-harmonic ripple in the DC-link voltage and torque in machine-based systems. Therefore, a dedicated control strategy for unbalanced operation is mandatory for reliable FRT performance of solar inverters.
The key is to independently control the positive and negative sequence currents. First, the grid voltage must be accurately decomposed into its positive and negative sequence components. Techniques like the Decoupled Double Synchronous Reference Frame (DDSRF) or methods based on delayed signal cancellation are highly effective. Once separated, the voltages and currents can be expressed in their respective positive (p) and negative (n) sequence dq-frames:
$$ \mathbf{u}_{dq} = \mathbf{u}_{dq}^p + \mathbf{u}_{dq}^n e^{j(\theta_n – \theta_p)} $$
$$ \mathbf{i}_{dq} = \mathbf{i}_{dq}^p + \mathbf{i}_{dq}^n e^{j(\theta_n – \theta_p)} $$
where $\theta_p$ and $\theta_n$ are the angles of the positive and negative sequence rotating frames, with $\theta_p – \theta_n = 2\omega t$.
The instantaneous active $(p)$ and reactive $(q)$ power can be derived from these components:
$$ p = P_0 + P_{c2} \cos(2\omega t) + P_{s2} \sin(2\omega t) $$
$$ q = Q_0 + Q_{c2} \cos(2\omega t) + Q_{s2} \sin(2\omega t) $$
The terms $P_0$ and $Q_0$ are the constant (average) active and reactive power components, while the terms with coefficients $P_{c2}$, $P_{s2}$, $Q_{c2}$, $Q_{s2}$ represent the oscillating power at twice the grid frequency. These oscillations are the direct cause of DC-link voltage ripple.
The proposed improved dual-sequence control strategy has multiple objectives: 1) Provide the required dynamic reactive support based on the positive-sequence voltage dip, 2) Manage active power to maintain DC-link stability and current limits, and 3) Eliminate negative-sequence current injection to ensure balanced output currents. This is achieved by setting the negative-sequence current references to zero $(I_d^{n*} = I_q^{n*} = 0)$ and calculating the positive-sequence references to deliver the desired average active power $(P_0^*)$ and reactive power $(Q_0^*)$:
$$ I_d^{p*} = \frac{P_0^* U_d^p + Q_0^* U_q^p}{(U_d^p)^2 + (U_q^p)^2} $$
$$ I_q^{p*} = \frac{P_0^* U_q^p – Q_0^* U_d^p}{(U_d^p)^2 + (U_q^p)^2} $$
Here, $P_0^*$ is derived from the active power management block (which may curtail power based on total current limits), and $Q_0^*$ is determined by the FRT requirement based on the positive-sequence voltage magnitude. Independent PI controllers in the positive and negative sequence dq-frames then regulate these currents. The controller outputs are the required inverter voltage references:
$$ V_d^{p*} = \left(K_p + \frac{K_i}{s}\right)(I_d^{p*} – I_d^p) + U_d^p – \omega L I_q^p $$
$$ V_q^{p*} = \left(K_p + \frac{K_i}{s}\right)(I_q^{p*} – I_q^p) + U_q^p + \omega L I_d^p $$
$$ V_d^{n*} = \left(K_p + \frac{K_i}{s}\right)(I_d^{n*} – I_d^n) + U_d^n – \omega L I_q^n $$
$$ V_q^{n*} = \left(K_p + \frac{K_i}{s}\right)(I_q^{n*} – I_q^n) + U_q^n + \omega L I_d^n $$
These voltage references are then transformed back to the stationary abc-frame to generate PWM signals for the three-level solar inverter.
| Power Component | Description | Primary Source | Control Objective |
|---|---|---|---|
| $P_0$ | Constant Active Power | Positive-seq. currents | Set by power management |
| $Q_0$ | Constant Reactive Power | Positive-seq. currents | Set by FRT voltage support |
| $P_{c2}, P_{s2}$ | Oscillating Active Power | Cross-coupling of pos./neg. seq. | Minimize via neg.-seq. current suppression |
| $Q_{c2}, Q_{s2}$ | Oscillating Reactive Power | Cross-coupling of pos./neg. seq. | Minimize via neg.-seq. current suppression |
To validate the proposed control strategies, detailed simulations of a 10 kW three-level solar inverter system were conducted. Key parameters include a DC-link voltage range of 430-900 V, grid line voltage of 400 V RMS, a switching frequency of 8 kHz, and an output filter inductance of 1.2 mH.
For a symmetrical fault with a 25% voltage dip, the FRT controller successfully injects the mandated reactive current $(I_q = 10 A$, which is 50% of a 20 A rated phase current). The current waveform shows the instantaneous transition from unity power factor operation to a state with significant reactive component. The active power is simultaneously curtailed to keep the total current magnitude at or below the rated value. Upon voltage recovery, the reactive current support is maintained for the required hold time before returning to normal operation. In a frequency excursion event, where the grid frequency rises to 50.5 Hz, the active power is reduced according to the stipulated $f-P$ droop characteristic, demonstrating the integrated power management capability of the solar inverter.
The effectiveness of the improved dual-sequence controller is starkly evident during a severe single-phase fault (80% dip in one phase). A conventional controller fails, leading to extreme current distortion and DC-link instability. In contrast, the proposed controller maintains balanced three-phase output currents with minimal low-order harmonics. While a second-harmonic ripple in the DC-link voltage is inevitable due to the oscillating power terms $P_{c2}$ and $P_{s2}$—which cannot be entirely eliminated while controlling constant power and zero negative-sequence current—its magnitude is contained within safe limits, allowing the solar inverter to remain connected and stable.
In conclusion, compliance with modern medium-voltage grid codes requires solar inverters to be equipped with sophisticated, multi-objective control algorithms. For symmetrical faults, a direct power control approach coordinating dynamic reactive current injection $(I_q)$ with active power curtailment $(P)$ is essential. This ensures DC-link voltage stability and prevents overcurrent while providing crucial voltage support to the grid. For the more challenging asymmetrical faults, an improved dual-sequence current control strategy is indispensable. By suppressing negative-sequence currents and calculating positive-sequence references based on FRT power requirements, solar inverters can achieve stable operation with acceptable current quality, fulfilling the fault ride-through mandate. These advanced control strategies transform solar inverters from mere power sources into active grid-supporting assets, enhancing the resilience and reliability of power networks with high renewable penetration. The continuous evolution of these controls will be paramount as the role of solar inverters in grid stability continues to expand.