The integration of distributed generation systems, particularly those based on photovoltaic (PV) technology, into the electrical grid presents both opportunities and challenges for modern power system operation. The inherent characteristics of **solar inverters**, such as fast response, proximity to loads, and potential for power quality regulation, make them valuable assets for grid support. This has been underscored by modern grid paradigms like the Smart Grid. Consequently, grid operators worldwide are formulating stringent technical guidelines for grid connection to ensure stability and reliability. Among these, the German Medium-Voltage Directive (MVD) by the BDEW is a prominent example, mandating advanced functionalities like reactive power compensation, fault ride-through (FRT), and active power management for **solar inverters** connecting to the medium-voltage network. This article delves into comprehensive control strategies enabling three-phase, three-level **solar inverters** to fulfill these critical requirements during both symmetrical and asymmetrical grid faults, ensuring stable operation and continued support to the grid.
The evolution of grid codes for **solar inverters** marks a shift from simple disconnection during disturbances to an active support role. The German MVD exemplifies this shift, imposing specific behaviors during voltage sags, swells, and frequency deviations. For symmetrical faults, the directive requires **solar inverters** to provide dynamic voltage support by injecting reactive current proportional to the voltage dip and to manage active power output to prevent overcurrent and DC-link overvoltage. For asymmetrical faults, while the directive’s specifics may vary, the general requirement is to remain connected without causing excessive current distortion or instability. Failure to meet these FRT requirements can lead to widespread tripping of generation, exacerbating the initial disturbance and potentially leading to cascading failures. Therefore, developing robust control algorithms for **solar inverters** is paramount for the secure integration of high-penetration PV systems.
The core challenge during grid faults is power imbalance. The grid side may not be able to accept the full power generated by the PV array, leading to a surge in the DC-link voltage of the **solar inverter** if the power is not properly managed. Simultaneously, the current controllers may saturate trying to push current into a faulted grid, leading to loss of control and potential damage. Therefore, the control system of a modern **solar inverter** must integrate several layers: maximum power point tracking (MPPT), DC-link voltage regulation, grid-synchronization, current control, and superimposed grid-support functions. The following sections detail a holistic control approach for a two-stage PV system (comprising a DC-DC boost stage and a three-level voltage source inverter) to achieve seamless FRT operation.
Technical Requirements of the Medium-Voltage Grid Directive
The German MVD outlines clear operational boundaries and required responses for distributed generators like those using **solar inverters**. The key requirements relevant to FRT control are summarized in the table below.
| Functionality | Steady-State Requirement | Transient (FRT) Requirement |
|---|---|---|
| Reactive Power / Power Factor (PF) Control | Must operate within PF = ±0.95 or follow a remote dispatch signal for reactive power (Q). | Must inject reactive current (Iq) during voltage dips. For a voltage drop (ΔU) > 10%, require Iq = 2% * In per 1% ΔU, up to 100% In. Response time < 20 ms. Support must be maintained for ≥500 ms after voltage recovery. |
| Active Power (P) Management | Must be capable of reducing P output based on a remote dispatch signal. | Must reduce P output during over-frequency events (f > 50.2 Hz). Reduction is linear, reaching 40% P reduction at 51.5 Hz. Must also reduce P during deep voltage sags to prevent overcurrent. |
| Fault Ride-Through | N/A | Must not disconnect for specified voltage dip/swell profiles (Low/High Voltage Ride-Through). Must remain stable and controlled during asymmetrical faults. |
| Short-Circuit Current | Must not exceed specified limits. | N/A (Inherently limited by PV panel characteristics in **solar inverters**). |
System Architecture of a Three-Level Solar Inverter
The considered system, as shown conceptually below, features a two-stage topology. The first stage is a DC-DC boost converter responsible for extracting maximum power from the PV array via an MPPT algorithm. Its output establishes the high-voltage DC-link. The second stage is a three-phase, three-level Neutral-Point-Clamped (NPC) voltage source inverter, which interfaces with the medium-voltage grid through an L-filter. This topology for **solar inverters** offers advantages like reduced switching losses and lower harmonic distortion compared to two-level inverters. The control system’s objectives are to regulate the DC-link voltage, control the injected grid current, and implement the advanced grid-support functions mandated by the MVD.
The mathematical model of the grid-connected inverter in the synchronous rotating reference frame (dq-frame) locked to the grid voltage fundamental frequency is foundational. The voltage and current dynamics are given by:
$$ L \frac{d\boldsymbol{i}_{dq}}{dt} = \boldsymbol{v}_{dq}^{inv} – \boldsymbol{v}_{dq}^{grid} – j\omega L \boldsymbol{i}_{dq} $$
where $\boldsymbol{i}_{dq} = i_d + j i_q$ is the complex current vector, $\boldsymbol{v}_{dq}^{inv}$ is the inverter voltage vector, $\boldsymbol{v}_{dq}^{grid} = v_d + j v_q$ is the grid voltage vector (with $v_d$ typically equal to the grid voltage magnitude and $v_q=0$ in normal operation), $L$ is the filter inductance, and $\omega$ is the grid angular frequency. The active ($P$) and reactive ($Q$) powers injected into the grid are:
$$ P = \frac{3}{2} (v_d i_d + v_q i_q) $$
$$ Q = \frac{3}{2} (v_q i_d – v_d i_q) $$
Under normal, balanced conditions with the d-axis aligned to the grid voltage vector ($v_d=V_{grid}, v_q=0$), these simplify to $P = \frac{3}{2} V_{grid} i_d$ and $Q = -\frac{3}{2} V_{grid} i_q$. This decoupling forms the basis for independent control of active and reactive power in **solar inverters**.
Control Strategy for Symmetrical Grid Faults
During a symmetrical three-phase voltage dip, the primary goals for the **solar inverter** are: 1) to provide the mandated reactive current support, and 2) to manage active power to keep the DC-link voltage stable and the output current within safe limits. These two objectives are deeply coupled and must be addressed simultaneously.
Integrated Reactive Power Compensation and Control
The control system must seamlessly switch between steady-state and transient reactive power modes. In steady-state, it should accept either a power factor (cosφ*) or a reactive power (Q_ref*) setpoint from a dispatch center. The current references are then calculated accordingly. For a given output current amplitude $I_{amp}^*$ (which is the output of the DC-link voltage controller) and a desired power factor, the dq-current references are:
$$ i_d^* = I_{amp}^* \cos\phi^* $$
$$ i_q^* = I_{amp}^* \sin(\arccos(\cos\phi^*)) $$
Alternatively, for a direct reactive power command $Q_{ref}^*$, the references become:
$$ i_q^* = \frac{2}{3} \frac{Q_{ref}^*}{v_d} $$
$$ i_d^* = \sqrt{(I_{amp}^*)^2 – (i_q^*)^2} $$
This last equation highlights the coupling: increasing reactive current $i_q$ for support necessarily reduces the available capacity for active current $i_d$ if the total current magnitude $I_{amp}^*$ is at its limit. This leads directly to the need for active power management.
During a transient fault, the priority shifts to dynamic voltage support. The required reactive current $I_{q,FRT}^*$ is determined instantly from the measured voltage dip magnitude $\Delta U$ (in per-unit of nominal voltage $U_n$):
$$ I_{q,FRT}^* = \begin{cases}
0 & \text{if } \Delta U \le 0.1 \\
k \cdot \Delta U \cdot I_n & \text{if } 0.1 < \Delta U \le 0.5 \\
1.0 \cdot I_n & \text{if } \Delta U > 0.5
\end{cases} $$
where $k$ is the gain factor (typically 2.0, meaning 2% of rated current per 1% voltage dip), and $I_n$ is the rated current of the **solar inverter**. This $I_{q,FRT}^*$ overrides any steady-state $i_q^*$ command. The control block for generating the final dq-current references $i_d^{**}$ and $i_q^{**}$ during this integrated operation is shown below.
Active Power Management and DC-Link Stability
The power balance equation at the DC-link of the **solar inverter** is crucial:
$$ P_{PV} – P_{out} = \frac{d}{dt} \left( \frac{1}{2} C_{dc} V_{dc}^2 \right) $$
where $P_{PV}$ is the power from the PV array (via the boost stage), $P_{out}$ is the power output to the grid, and $C_{dc}$ is the DC-link capacitance. During a fault, $P_{out}$ may be limited because the grid voltage is low ($P_{out} \propto V_{grid} \cdot i_d$). If $P_{PV}$ remains at its MPPT value, the excess power charges the capacitor, causing $V_{dc}$ to rise dangerously. Therefore, $P_{PV}$ must be reduced (derated) to match the deliverable $P_{out}$.
This is achieved by overriding the MPPT algorithm. A power reference $P_{ref}$ is generated based on two possible conditions: 1) A dispatch command, or 2) A transient requirement. The most critical transient requirement comes from the current limit. The total output current amplitude must not exceed the **solar inverter’s** rated value $I_{n,amp}$:
$$ I_{amp} = \sqrt{i_d^2 + i_q^2} \le I_{n,amp} $$
Given the reactive current requirement $i_q^* = I_{q,FRT}^*$, the maximum allowable active current is:
$$ i_{d,max}^* = \sqrt{ (I_{n,amp})^2 – (I_{q,FRT}^*)^2 } $$
The corresponding maximum allowable active power output is then $P_{max} = \frac{3}{2} v_d \cdot i_{d,max}^*$. This $P_{max}$ becomes the active power reference $P_{ref}$ for derating. The derating algorithm acts on the front-end boost converter by shifting the PV array operating point away from the MPPT point towards a higher voltage, lower current point, thereby reducing $P_{PV}$. This power management loop is essential for the successful fault ride-through of **solar inverters**.
Additionally, the MVD requires frequency-based active power derating. For grid frequencies $f$ above 50.2 Hz, the required power reduction $\Delta P$ is:
$$ \Delta P = 20 \cdot P_m \cdot \frac{50.2 – f}{50.0} \quad \text{for} \quad 50.2 < f \le 51.5 \text{ Hz} $$
where $P_m$ is the available power at the time. This generates another potential $P_{ref}$ for derating.
The complete control structure for the **solar inverter** during symmetrical faults thus integrates a DC-link voltage PI controller producing $I_{amp}^*$, a reactive current calculator producing $I_{q,FRT}^*$, a current limiter that computes $i_{d}^{**}$ and $i_{q}^{**}$, and inner PI-based current regulators in the dq-frame. The output of these regulators, after decoupling and feedforward terms, generates the modulating signals for the three-level PWM.
Control Strategy for Asymmetrical Grid Faults
Asymmetrical faults (e.g., single-phase or phase-to-phase faults) are more complex because they lead to unbalanced grid voltages containing both positive-sequence (fundamental) and negative-sequence components. If a standard synchronous frame controller (locked to the positive-sequence) is used, the negative-sequence voltage appears as a 100 Hz (2ω) disturbance, causing severe 100 Hz oscillations in the measured dq-currents and, consequently, in the injected power and DC-link voltage. This can destabilize the **solar inverter** and produce highly distorted output currents, failing power quality standards.
Modeling in Unbalanced Conditions
Under unbalanced grid voltages, the voltage and current vectors in the positive-sequence synchronous reference frame (denoted by superscript ‘p’) can be expressed as:
$$ \boldsymbol{v}_{dq} = \boldsymbol{v}_{dq}^p + \boldsymbol{v}_{dq}^n e^{j(\theta_n – \theta_p)} = \boldsymbol{v}_{dq}^p + \boldsymbol{v}_{dq}^n e^{-j2\omega t} $$
$$ \boldsymbol{i}_{dq} = \boldsymbol{i}_{dq}^p + \boldsymbol{i}_{dq}^n e^{j(\theta_n – \theta_p)} = \boldsymbol{i}_{dq}^p + \boldsymbol{i}_{dq}^n e^{-j2\omega t} $$
where $\boldsymbol{v}_{dq}^n$ and $\boldsymbol{i}_{dq}^n$ are the negative-sequence vectors in their own negative-sequence rotating frame, and $\theta_p – \theta_n = 2\omega t$. The instantaneous active and reactive powers become:
$$ p(t) = P_0 + P_{c2} \cos(2\omega t) + P_{s2} \sin(2\omega t) $$
$$ q(t) = Q_0 + Q_{c2} \cos(2\omega t) + Q_{s2} \sin(2\omega t) $$
where $P_0$ and $Q_0$ are the average active and reactive power components, and the terms with $\cos(2\omega t)$ and $\sin(2\omega t)$ represent the oscillating power components at double the grid frequency.
Improved Dual-Sequence Current Control with Power Management
The goal for the **solar inverter** during asymmetrical faults is to maintain stable operation, provide grid support, and often to eliminate the negative-sequence current injection to avoid overheating other equipment. An improved dual-sequence current control strategy is employed. It uses a decoupled double synchronous reference frame (DDSRF) phase-locked loop (PLL) to accurately extract the positive-sequence ($v_d^p$, $v_q^p$) and negative-sequence ($v_d^n$, $v_q^n$) voltage components.
Separate current controllers are designed for the positive-sequence and negative-sequence currents. The key improvement lies in calculating the positive-sequence current references ($i_d^{p*}$, $i_q^{p*}$) to achieve specific power objectives while setting the negative-sequence references to zero ($i_d^{n*}=0$, $i_q^{n*}=0$) to suppress negative-sequence current injection.
To simultaneously deliver an average active power $P_0^*$ (which may be derated due to current limits as in the symmetrical case) and an average reactive power $Q_0^*$ (for voltage support), the positive-sequence current references are derived from the power equations:
$$ P_0^* = \frac{3}{2}(v_d^p i_d^p + v_q^p i_q^p) $$
$$ Q_0^* = \frac{3}{2}(v_q^p i_d^p – v_d^p i_q^p) $$
Solving this system for $i_d^{p*}$ and $i_q^{p*}$ yields:
$$ i_d^{p*} = \frac{2}{3} \frac{P_0^* v_d^p + Q_0^* v_q^p}{(v_d^p)^2 + (v_q^p)^2} $$
$$ i_q^{p*} = \frac{2}{3} \frac{P_0^* v_q^p – Q_0^* v_d^p}{(v_d^p)^2 + (v_q^p)^2} $$
The average active power reference $P_0^*$ is determined by the higher-level power management layer, considering the MPPT power, current limits, and frequency-watt requirements. $Q_0^*$ is determined by the voltage support requirement, typically based on the magnitude of the positive-sequence voltage dip $\Delta U^p$.
With the references defined, independent PI controllers regulate the positive-sequence and negative-sequence currents. The outputs of these controllers, after adding appropriate decoupling and feedforward terms, are combined and transformed back to the stationary abc frame to generate the PWM signals for the **solar inverter**. The control laws are:
For the positive-sequence controller:
$$ u_d^{p*} = \left( K_p + \frac{K_i}{s} \right)(i_d^{p*} – i_d^p) + v_d^p – \omega L i_q^p $$
$$ u_q^{p*} = \left( K_p + \frac{K_i}{s} \right)(i_q^{p*} – i_q^p) + v_q^p + \omega L i_d^p $$
For the negative-sequence controller:
$$ u_d^{n*} = \left( K_p + \frac{K_i}{s} \right)(i_d^{n*} – i_d^n) + v_d^n + \omega L i_q^n $$
$$ u_q^{n*} = \left( K_p + \frac{K_i}{s} \right)(i_q^{n*} – i_q^n) + v_q^n – \omega L i_d^n $$
The final modulating voltage in the positive-sequence dq-frame is: $\boldsymbol{u}_{dq}^{mod} = \boldsymbol{u}_{dq}^{p*} + \boldsymbol{u}_{dq}^{n*} e^{-j2\omega t}$. This strategy allows the **solar inverter** to maintain a balanced, sinusoidal current output despite severe grid unbalance, achieving stable fault ride-through.
Simulation and Performance Analysis
To validate the proposed strategies, a detailed simulation model of a 10 kW three-level **solar inverter** system was developed with the following parameters: DC-link voltage range 430-900 V, grid line voltage 400 V RMS, rated current 20 A peak, switching frequency 8 kHz, output filter L=1.2 mH.
Case 1: Symmetrical Voltage Dip of 15%. At t=0.08s, a balanced three-phase voltage dip of 15% occurs. The controller immediately calculates the required reactive current: $I_{q,FRT}^* = 2.0 \times 0.15 \times 20A = 6A$. The simulation shows the $i_q$ component rising to 6A within 5ms, meeting the <20ms response mandate. The total output current initially exceeds 20A due to the added reactive component. The power management layer then activates, derating the active power by reducing $i_d$, bringing the total current magnitude back to or below 20A. The DC-link voltage remains stable throughout the event.
Case 2: Symmetrical Voltage Dip of 25% with Recovery. At t=0.08s, a 25% dip occurs, demanding $I_{q,FRT}^* = 10A$. This is correctly injected. At t=0.15s, the voltage recovers to 90%. According to the MVD, the reactive support is maintained for 500ms. During this hold period, as the voltage is higher, the same $i_q$ requires less voltage from the inverter, and the active current $i_d$ can slightly increase, demonstrating the dynamic interplay of the controllers in the **solar inverter**.
Case 3: Frequency-based Derating. With the grid frequency rising to 50.5 Hz at t=0.08s, the required power reduction is calculated as $\Delta P = 20 \times 10kW \times (50.2-50.5)/50.0 = -1.2kW$, setting $P_{ref}=8.8kW$. The MPPT algorithm is derated accordingly, and the output current amplitude reduces proportionally, demonstrating compliance with the frequency-watt requirement.
Case 4: Severe Single-Phase Voltage Dip (80% Asymmetrical Fault). A standard single dq-frame controller fails, causing current runaway and tripping. With the improved dual-sequence controller enabled, the system remains stable. The positive-sequence voltage dip is calculated, and appropriate $P_0^*$ (derated) and $Q_0^*$ (support) references are generated. The negative-sequence current references are set to zero. The result is that the **solar inverter** injects balanced, sinusoidal three-phase currents with minimal negative-sequence components, and the DC-link voltage exhibits only a small, well-damped 100 Hz ripple, confirming successful ride-through.
Conclusion
The successful integration of large-scale PV generation into modern medium-voltage grids necessitates that **solar inverters** evolve from simple power converters into intelligent grid assets. This article has presented a comprehensive set of advanced control strategies enabling three-phase **solar inverters** to meet rigorous grid codes like the German MVD. The integrated approach combines dynamic reactive current injection for voltage support with active power management for stability during symmetrical faults. For the more challenging asymmetrical fault conditions, an improved dual-sequence current control strategy is essential to suppress negative-sequence currents and maintain high-quality power injection. These strategies collectively ensure that **solar inverters** can ride through both balanced and unbalanced grid disturbances, actively supporting network voltage and frequency, and thereby enhancing the overall resilience and reliability of the power system. The implementation of such algorithms is now a fundamental requirement for the next generation of grid-supportive **solar inverters**.