The global energy landscape is undergoing a profound transformation, driven by the imperative for sustainability and the rapid integration of renewable energy sources. Among these, solar photovoltaic (PV) power generation stands out due to its scalability and decreasing cost. A significant portion of solar capacity is now deployed as distributed generation (DG) systems, feeding power directly into the distribution network. The core interface between a PV array and the utility grid is the solar inverter. Its primary function is to convert the DC power from the panels into grid-compliant AC power. However, the stability and reliability of modern power grids, with their increasing penetration of inverter-based resources, hinge critically on the dynamic behavior of these solar inverters during grid disturbances, particularly voltage dips or sags.
Voltage sags are among the most common power quality issues, caused by faults, large motor starting, or other network events. Traditionally, generators were required to remain connected and support the grid during such faults. With the shift to inverter-based generation, this requirement has been formalized into grid codes mandating Low Voltage Ride-Through (LVRT) capability. An inverter with LVRT must not disconnect during specified voltage dips; instead, it must remain synchronized and often provide reactive current support to help stabilize the grid voltage. The challenge for solar inverters lies in their inherent characteristics: the power output is stochastic, dependent on irradiance, and they lack the natural rotational inertia of synchronous machines. This randomness can exacerbate frequency fluctuations during faults, adversely affecting the inverter’s dynamic response during LVRT, potentially leading to nuisance tripping and reduced system stability.
This article delves into an advanced control methodology designed to ensure robust LVRT performance for solar inverters in distributed generation systems. The approach focuses on three pillars: fast and accurate detection of voltage sags, precise control of the inverter’s output current during the fault, and the effective mitigation of negative-sequence components caused by asymmetrical faults, all while maintaining stable operation.
1. LVRT Requirements and the Core Challenge
Grid codes worldwide define specific LVRT profiles that generation plants must adhere to. A typical requirement is illustrated in the figure below. The inverter must stay connected for faults where the voltage at the Point of Common Coupling (PCC) remains above the curve for a defined duration. Crucially, during the fault, the inverter is often required to inject reactive current (Iq) proportional to the voltage drop. The general requirement can be summarized as:
$$ I_{q}^{ref} \geq 1.5 I_N (0.9 – U_T), \quad \text{for} \quad 0.2 \leq U_T \leq 0.9 $$
$$ I_{q}^{ref} \geq 1.5 I_N, \quad \text{for} \quad U_T \leq 0.2 $$
$$ I_{q}^{ref} = 0, \quad \text{for} \quad U_T > 0.9 $$
where \( I_{q}^{ref} \) is the reference reactive current, \( I_N \) is the rated current of the solar inverter, and \( U_T \) is the per-unit voltage at the PCC.
The primary technical challenge is twofold: First, to accurately and swiftly detect the depth and phase of the voltage sag, especially under unbalanced conditions which create negative-sequence components. Second, to control the inverter’s current controllers to meet the reactive current demand without exceeding the maximum current capability of the power semiconductors, which is governed by:
$$ k I_N \geq \sqrt{(I_{q}^{ref})^2 + (I_{d}^{ref})^2} $$
where \( I_{d}^{ref} \) is the active current reference and \( k \) is the overcurrent capability factor (e.g., 1.1 to 1.2).
2. System Modeling and Voltage Sag Detection
Accurate control begins with precise measurement. We model the grid-connected solar inverter system in the synchronous rotating reference frame (d-q frame), which simplifies the control of AC quantities by transforming them into DC values. The fundamental voltage equation for the inverter output filter is:
$$ \mathbf{U}_{dq} = L \frac{d\mathbf{I}_{dq}}{dt} + \mathbf{e}_{dq} $$
where \( \mathbf{U}_{dq} = [U_d, U_q]^T \) is the inverter output voltage vector, \( \mathbf{I}_{dq} = [I_d, I_q]^T \) is the grid current vector, \( \mathbf{e}_{dq} = [e_d, e_q]^T \) is the grid voltage vector, and \( L \) is the filter inductance.
To enable fast response, a predictive model can be used to estimate the future current:
$$ \mathbf{I}_{dq}(t+1) = \mathbf{I}_{dq}(t) + \frac{T_s}{L} [\mathbf{U}_{dq}(t) – \mathbf{e}_{dq}(t)] $$
where \( T_s \) is the sampling period.
The three-phase grid voltages (\( u_a, u_b, u_c \)) are measured and contain fundamental and harmonic components. Using Clarke (\( \alpha\beta \)) and Park (dq) transformations, we extract the fundamental positive-sequence component in the d-q frame. Under ideal balanced conditions, this results in constant d and q components: \( u_d = \sqrt{3}U_{m1}\cos\theta \), \( u_q = -\sqrt{3}U_{m1}\sin\theta \), where \( U_{m1} \) is the peak phase voltage and \( \theta \) is the phase angle.
However, during an asymmetrical fault (e.g., a single-phase sag), the voltage contains both positive and negative sequence components. The transformed voltage in the d-q frame becomes:
$$ \begin{bmatrix} u_d \\ u_q \end{bmatrix} = \begin{bmatrix} U_{m1}^+ \\ 0 \end{bmatrix} + \begin{bmatrix} U_{m1}^- \cos(-2\omega t + \phi^-) \\ U_{m1}^- \sin(-2\omega t + \phi^-) \end{bmatrix} $$
where the superscripts \( + \) and \( – \) denote positive and negative sequence, \( \omega \) is the grid frequency, and \( \phi \) is the phase difference. The negative sequence manifests as a ripple at twice the grid frequency (2ω) in the d-q frame. By applying a suitable filter (like a Decoupled Double Synchronous Reference Frame PLL), the positive sequence component \( U_{m1}^+ \) can be cleanly extracted. The magnitude of this component, compared to its rated value, directly gives the per-unit voltage \( U_T \) used in the LVRT requirement. This method allows for rapid and accurate detection of both symmetrical and asymmetrical voltage sags.
3. Current Control Strategy with Transient Impedance Reshaping
Once a sag is detected, the current references must be generated according to the grid code. The active power/current reference is typically reduced to limit the power imbalance that causes DC-link voltage rise. Priority is given to reactive current injection. The challenge for the inner current control loops is to track these new references accurately and quickly without overshoot or instability, which is exacerbated by the changed grid impedance during a fault.
To address this, we employ an impedance reshaping technique. The inherent output impedance of the solar inverter, primarily inductive from the filter, interacts with the grid impedance. This interaction affects stability margins. The proposed method uses a compensator \( G(s) \) to reshape the inverter’s transient output impedance, improving its robustness during the fault transient. The current control law incorporating this is derived as:
$$ \mathbf{I}_{ref} = \frac{G(s) \cdot (\mathbf{e}^* – \mathbf{u}_o)}{sL_m + G(s)R_m} $$
where \( \mathbf{I}_{ref} \) is the final current reference signal, \( \mathbf{e}^* \) is the voltage command, \( \mathbf{u}_o \) is the measured PCC voltage, and \( L_m \) and \( R_m \) are virtual inductance and resistance terms used for active damping and shaping the output characteristic.
The compensator \( G(s) \) is designed as a lead-lag network with a first-order filter for high-frequency noise attenuation:
$$ G(s) = \frac{K_d (1 + T_d s)}{(\zeta T_d s + 1)(\tau s + 1)} $$
Here, \( K_d \) is the gain, \( T_d \) the lead time constant, \( \zeta \) the lag coefficient, and \( \tau \) the inertia constant. By carefully tuning \( G(s) \), the inverter’s impedance in the critical frequency range (around the control bandwidth) is modified to ensure sufficient phase margin and damping even when the grid impedance changes due to the fault. This results in a stable, well-damped transient response when the current references switch to LVRT mode.
4. Negative-Sequence Compensation and Voltage Modulation
For unbalanced sags, the negative-sequence voltage component can cause severe problems like unbalanced currents, increased torque pulsations in nearby motors, and overheating. Advanced solar inverters can be controlled to mitigate this. Based on instantaneous power theory, separate control loops for the positive and negative sequence currents can be established. The goal during LVRT is often to eliminate the negative-sequence current to prevent equipment stress, or to inject negative-sequence reactive current to help balance the voltages.
The core of this implementation lies in the modulation stage. A Space Vector Pulse Width Modulation (SVPWM) algorithm is highly effective. The current controllers generate reference voltage signals in the positive and negative sequence d-q frames (\( u_d^{p*}, u_q^{p*}, u_d^{n*}, u_q^{n*} \)). These are transformed back to the three-phase stationary frame to create the reference voltage waveform. The SVPWM algorithm then calculates the appropriate switching states and duty cycles for the inverter’s power switches to synthesize this reference. The mathematical representation for the positive-sequence control voltage is given by:
$$ u_d^{p*} = (k_{ip} + \frac{k_{ii}}{s})(i_d^{p*} – i_d^{p}) – \omega L’ i_q^{p} + \hat{e}_d^{p} $$
$$ u_q^{p*} = (k_{ip} + \frac{k_{ii}}{s})(i_q^{p*} – i_q^{p}) + \omega L’ i_d^{p} + \hat{e}_q^{p} $$
where \( k_{ip}, k_{ii} \) are the PI controller gains, \( \omega \) is the grid frequency, \( L’ \) is the estimated filter inductance, and \( \hat{e}_d^{p}, \hat{e}_q^{p} \) are feedforward terms of the estimated positive-sequence grid voltage. A similar, decoupled structure is used for the negative-sequence controller, but operating at the negative synchronous frequency (-ω). By independently controlling these components, the solar inverter can dynamically adjust its output to counteract the grid voltage negative-sequence component, enhancing the overall power quality and stability during the LVRT event.
5. Method Comparison and Performance Analysis
To contextualize the proposed integrated approach (combining fast sequence separation, impedance-reshaped current control, and dual-sequence SVPWM), it is useful to compare it with other common LVRT strategies for solar inverters.
| Methodology | Core Principle | Advantages | Limitations / Challenges |
|---|---|---|---|
| Power / DC-Link Control | Reduces active power reference during sag to limit DC-link overvoltage. | Simple logic, protects inverter hardware. | Does not actively support grid voltage; may not meet reactive current injection mandates. |
| Reactive Current Injection (Basic) | Injects reactive current proportional to voltage dip. | Directly addresses grid code requirements. | Performance sensitive to detection speed/accuracy; risk of current overshoot if not properly controlled. |
| Virtual Synchronous Generator (VSG) | Mimics inertia/damping of a synchronous machine via control algorithms. | Provides inherent frequency support, improves system stability. | Complex control; LVRT dynamic performance may be constrained by virtual inertia parameters. |
| Proposed Integrated Approach | Combines sequence-based detection, impedance-reshaped current control, and dual-sequence compensation. | Fast, accurate response to symmetrical/asymmetrical faults; stable transient; actively mitigates negative sequence. | Higher computational and tuning complexity. |
The performance of the proposed strategy can be evaluated through simulation of a representative system. Key simulation parameters for a 100kW two-stage solar inverter system are summarized below:
| Parameter | Value |
|---|---|
| Rated System Power | 100 kW |
| DC-Link Voltage | 800 V |
| Grid Voltage (L-N, RMS) | 220 V |
| Grid Frequency | 50 Hz |
| Solar Inverter Rated Current | 200 A |
| Filter Inductance (Inverter-side) | 0.4 mH |
During a simulated asymmetrical voltage sag (e.g., 25% single-phase dip), the proposed control ensures:
- Fast Reactive Current Buildup: The d-axis (active) current is reduced, and the q-axis (reactive) current is rapidly increased to the value mandated by the LVRT curve, supporting the grid voltage.
- Stable Power Transition: The output power from the PV array smoothly transitions to a lower active power and higher reactive power operating point without instability.
- Superior Frequency Stability: A key metric is the frequency deviation at the PCC. Compared to other methods, the impedance-reshaping control maintains frequency fluctuations within a very tight band (e.g., ±0.5 Hz), whereas other strategies may exhibit wider swings (>2 Hz).
- Negative Sequence Suppression: The dual-sequence control effectively minimizes negative-sequence currents, preventing unbalanced loading and associated thermal stress on system components.
The improved frequency stability is a direct consequence of the well-damped current controller and the rapid, balanced reactive support provided by the solar inverter, which helps to stabilize the local grid voltage and, by extension, the frequency.
6. Conclusion
The successful integration of high-penetration distributed solar generation necessitates that solar inverters evolve from simple power converters to intelligent grid-supporting assets. Robust Low Voltage Ride-Through capability is a cornerstone of this evolution. The integrated technical approach discussed herein—combining precise sequence-component-based voltage detection, a robust current control strategy enhanced by transient impedance reshaping, and advanced modulation with negative-sequence compensation—provides a comprehensive solution. This methodology ensures that solar inverters can not only remain connected during grid faults but also actively contribute to system stability by providing fast and well-damped reactive current support and improving power quality. This enhances the overall resilience of the power grid. While the approach demonstrates superior performance in defined scenarios, ongoing work focuses on optimizing its adaptability and performance across the full spectrum of real-world grid conditions and fault types. The continuous advancement of such control strategies is essential for building a secure, reliable, and sustainable electricity grid powered significantly by solar inverters and other renewable resources.
7. Summary of Key Techniques
| Functional Block | Technique Employed | Key Mathematical Expression / Purpose |
|---|---|---|
| Sag Detection | Decoupled Double Synchronous Reference Frame PLL | $$ \begin{bmatrix} u_d \\ u_q \end{bmatrix} = \begin{bmatrix} U_{m1}^+ \\ 0 \end{bmatrix} + \begin{bmatrix} U_{m1}^- \cos(-2\omega t + \phi^-) \\ U_{m1}^- \sin(-2\omega t + \phi^-) \end{bmatrix} $$ Extracts clean positive-sequence voltage (\(U_T\)) for LVRT logic. |
| Current Reference Generation | Grid-Code Compliant Reactive Priority | $$ I_{q}^{ref} \geq 1.5 I_N (0.9 – U_T) $$ $$ k I_N \geq \sqrt{(I_{q}^{ref})^2 + (I_{d}^{ref})^2} $$ Ensures compliance while protecting the solar inverter. |
| Inner Current Control | Impedance Reshaping via Compensator | $$ G(s) = \frac{K_d (1 + T_d s)}{(\zeta T_d s + 1)(\tau s + 1)} $$ $$ \mathbf{I}_{ref} = \frac{G(s) \cdot (\mathbf{e}^* – \mathbf{u}_o)}{sL_m + G(s)R_m} $$ Enhances transient stability and damping during fault. |
| Modulation & Compensation | Dual-Sequence SVPWM Control | $$ u_d^{p*} = (k_{ip} + \frac{k_{ii}}{s})(i_d^{p*} – i_d^{p}) – \omega L’ i_q^{p} + \hat{e}_d^{p} $$ Independent control of positive/negative sequences to counteract unbalanced voltages. |