The global energy landscape is undergoing a profound transformation, driven by the imperative to transition towards sustainable and low-carbon energy sources. In this context, photovoltaic (PV) power generation has emerged as a cornerstone technology. The increasing penetration of distributed PV systems, however, presents significant challenges to the stability and power quality of the electrical grid. Conventional grid-connected solar inverters primarily operate as current sources, relying entirely on the main grid to provide a stable voltage and frequency reference. This dependency creates a critical vulnerability: during grid faults or outages, these inverters are typically mandated to disconnect via anti-islanding protection, leading to a complete interruption of power supply to local loads. While integrating energy storage systems (ESS) can mitigate this issue by forming a stable islanded microgrid, the high cost and finite capacity of batteries remain practical constraints. Consequently, there is a pressing need to explore control strategies that enable solar inverters to provide autonomous voltage support even in the absence of both the main grid and energy storage buffers.
This article delves into the advanced control paradigms necessary for a solar inverter to function reliably across multiple operational modes. The core objective is to enable a seamless and robust transition between grid-connected (current-source) and storage-less islanded (voltage-source) operation, ensuring uninterrupted power to critical loads. We will dissect the system architecture, detail the specific control strategies for each mode, and formulate a novel seamless transfer strategy. The efficacy of the proposed approach will be validated through experimental results.
System Architecture and Operational Modes
The foundation of a multi-mode capable solar inverter lies in its power circuit topology and the logical definition of its states. The chosen topology must provide sufficient degrees of freedom for independent control of both DC-link voltage and AC output characteristics.
A two-stage configuration, comprising a front-end DC-DC converter and a rear-end inverter, is ideally suited for this application. This structure decouples the maximum power point tracking (MPPT) or power regulation of the PV panels from the AC voltage and frequency control tasks. For this study, we employ a two-level Boost converter as the DC-DC stage and a three-level Neutral Point Clamped (NPC) inverter as the AC stage. The NPC topology offers advantages such as reduced output voltage harmonic distortion, lower switching losses, and decreased voltage stress on the semiconductor devices compared to traditional two-level inverters, contributing to higher efficiency and power quality.
The system interfaces with the main grid through a static transfer switch (STS, K1). The operational logic revolves around four distinct modes, as summarized in the table below:
| Mode | Switch K1 State | Inverter Control Type | Voltage/Frequency Support | Primary Control Objective |
|---|---|---|---|---|
| Grid-Connected (GC) | Closed | Current-Source Control | Provided by Main Grid | Inject dispatched active/reactive power (Pref, Qref). |
| Storage-less Islanded (SLI) | Open | Voltage-Source Control (V/f) | Provided by the Solar Inverter itself | Maintain stable AC voltage amplitude and frequency for the local load. |
| Islanding Transition | Opening | Switching from Current to Voltage Control | Transitioning | Seamless handover from grid-following to grid-forming control without disruption. |
| Reconnection Transition | Pre-sync then Closing | Pre-synchronization then Switching to Current Control | Synchronizing with Grid | Align inverter output voltage with grid voltage before re-closing the switch. |
In the Grid-Connected (GC) Mode, the solar inverter acts as a controlled current source. It injects power according to references (Pref, Qref) provided by a higher-level energy management system. The stability of the DC-link voltage is managed by the front-end Boost converter. In the Storage-less Islanded (SLI) Mode, the solar inverter must transform into a grid-forming unit, establishing and regulating the AC bus voltage and frequency solely based on its internal control loops. The power balance between the variable PV generation and the local load is maintained by adjusting the PV operating point via the DC-link voltage control. The core challenge addressed in this work is the design of control algorithms for the smooth and transient-free transitions between these two fundamentally different modes.
Detailed Control Strategy for Each Mode
Grid-Connected (Current-Source) Control
In this mode, the solar inverter’s primary control objective is to accurately feed active and reactive power into the grid. Using a Phase-Locked Loop (PLL) synchronized to the grid voltage ($u_g$), the control is implemented in the synchronous reference frame (d-q frame), where the grid voltage vector is aligned with the d-axis ($u_{gd} = U_g, u_{gq} \approx 0$). The injected three-phase power is given by:
$$
\begin{aligned}
P &= \frac{3}{2} (u_{gd} i_{d} + u_{gq} i_{q}) \approx \frac{3}{2} U_g i_{d} \\
Q &= \frac{3}{2} (u_{gq} i_{d} – u_{gd} i_{q}) \approx -\frac{3}{2} U_g i_{q}
\end{aligned}
$$
where $i_d$ and $i_q$ are the d- and q-axis components of the inverter output current. Therefore, the current references for the inner current control loops are derived directly from the power references:
$$
\begin{aligned}
i_{d\_ref} &= \frac{2 P_{ref}}{3 U_g} \\
i_{q\_ref} &= -\frac{2 Q_{ref}}{3 U_g}
\end{aligned}
$$
The front-end Boost converter regulates the DC-link voltage ($u_{dc}$) to a fixed reference ($u_{dc\_ref}$). A cascaded control structure is used: an outer voltage PI controller generates the reference for the inductor current ($i_{L\_ref}$), which is then tracked by an inner current PI controller. This ensures that the power drawn from the PV array matches the power injected into the grid plus losses, thereby stabilizing the DC-link. The overall structure prioritizes power dispatch over MPPT, allowing the solar inverter to participate in grid support services. MPPT is engaged only when the available PV power is less than the commanded $P_{ref}$.
Storage-less Islanded (Voltage-Source) Control
Upon disconnection from the grid, the solar inverter must instantaneously assume responsibility for the local AC bus. The control strategy shifts to voltage-source or grid-forming control. The inverter’s output voltage amplitude and frequency are now dictated by internal references ($U_{ref}, f_{ref}$).
The critical task in this mode is maintaining real-time power balance. The load power demand ($P_{load}, Q_{load}$) must be matched exactly by the PV inverter’s output. Any imbalance causes the DC-link voltage to deviate. This deviation is used as the key feedback signal. An outer DC-link voltage PI controller adjusts the d-axis current reference ($i_{d\_ref}$) for the inner current control loop. Effectively, this controller moves the PV panel’s operating point on its I-V curve: if the load demand exceeds PV generation, $u_{dc}$ drops, the controller increases $i_{d\_ref}$, drawing more power from the PV array (moving it closer to or away from the MPP), and vice versa. The q-axis current reference ($i_{q\_ref}$) is generated by an AC voltage magnitude PI controller to regulate the output voltage. The control structure ensures that the solar inverter autonomously adjusts its output power to meet the load demand while maintaining stable voltage and frequency, all without any storage element.
Proposed Seamless Transfer Strategy
The abrupt change in control objectives and reference frames during a mode transition can cause severe voltage sags, frequency deviations, and overcurrent transients if not managed carefully. The proposed strategy ensures continuity by initializing the integrators of the new control loops with the final states of the previous loops.
Transition from Grid-Connected to Storage-less Islanded Mode
When a grid fault is detected (e.g., loss of voltage or frequency beyond limits), the static switch K1 is commanded to open. Simultaneously, the inverter control must switch from current-source to voltage-source control. The challenge is to avoid a discontinuity in the current reference ($i_{dq\_ref}$) fed to the inner current regulators, which are common to both modes.
Strategy: At the instant of transition ($t = t_{switch}$):
- The voltage controller (previously inactive in GC mode) is activated.
- The initial output of the voltage controller’s PI regulators (both for amplitude and DC-link) is forcefully initialized to the last value of the current references from the GC mode. Mathematically:
$$
\begin{aligned}
\text{PI}_{V\_out}(t_{switch}) &= i_{d\_ref\_GC}(t_{switch}^{-}) \\
\text{PI}_{Q\_out}(t_{switch}) &= i_{q\_ref\_GC}(t_{switch}^{-})
\end{aligned}
$$ - The phase angle generation for the voltage-source mode is initialized with the last phase angle from the PLL: $\theta_{V/f}(t_{switch}) = \theta_{PLL}(t_{switch}^{-})$.
This ensures that the current reference vector ($i_{dq\_ref}$) and the output voltage phase experience no step change. The inner current loop continues to track a smooth reference, while the newly activated outer voltage loops gradually take over regulation from their initialized states. The solar inverter smoothly transitions from injecting a specific current into the grid to forming a voltage for the islanded load.
Transition from Storage-less Islanded to Grid-Connected Mode (Reconnection)
Reconnecting to the grid requires careful synchronization to prevent large inrush currents. Before closing switch K1, the inverter’s output voltage must be synchronized in phase, frequency, and amplitude with the grid voltage.
Pre-synchronization Strategy: A dedicated pre-synchronization controller is activated.
- Phase/Frequency Synchronization: The difference between the grid PLL angle ($\theta_{PLL}$) and the inverter’s internal angle ($\theta_{V/f}$) is processed by a PI controller. Its output ($\Delta \omega$) is added to the nominal frequency reference to adjust the inverter’s frequency until the phase error is zero.
$$ \Delta \omega = K_{p\_\theta} (\theta_{PLL} – \theta_{V/f}) + K_{i\_\theta} \int (\theta_{PLL} – \theta_{V/f}) dt $$ - Voltage Amplitude Synchronization: The difference between the measured grid voltage ($U_g$) and the inverter’s voltage reference ($U_{ref}$) is processed by another PI controller. Its output ($\Delta U$) adjusts the inverter’s voltage magnitude reference.
$$ \Delta U = K_{p\_U} (U_g – U_{V/f}) + K_{i\_U} \int (U_g – U_{V/f}) dt $$
Once both phase and voltage errors are within a small tolerance band for a defined period, synchronization is complete. The static switch K1 is closed. At this precise moment, the control strategy is switched back to grid-connected current control. Crucially, the current references for the GC mode are initialized with the final output currents of the SLI mode to, again, prevent any reference jump and ensure a bumpless transfer. The solar inverter seamlessly starts injecting the dispatched power ($P_{ref}, Q_{ref}$) into the grid.
Experimental Validation
A laboratory-scale prototype was built to validate the proposed multi-mode control and seamless transfer strategy for the solar inverter. The setup consisted of a programmable DC source with a series resistor emulating a PV panel’s I-V characteristic, a two-stage solar inverter (Boost + NPC), an LCL filter, local resistive loads, and a back-to-back STS connection to the grid simulator.
Scenario 1: Grid Fault to Islanded Operation. Initially, the solar inverter operated in GC mode, injecting 210 W. A grid fault was simulated by opening the upstream breaker. The solar inverter detected the fault, opened its STS (K1), and executed the transition to SLI mode. Key waveforms are described:
- DC-link Voltage ($u_{dc}$): Experienced a minor, well-damped dip (less than 5%) and recovered swiftly to its reference value of 200V, demonstrating effective power balance control.
- PV Voltage ($u_{pv}$): Shifted from ~74V (GC operating point) to ~70V (new SLI operating point), showing the automatic adjustment of the PV generation point to match the load.
- AC Output Voltage ($u_m$) and Current ($i_m$): The transition was smooth, with no observable interruption, voltage distortion, or current surge. The load voltage remained sinusoidal and stable throughout.
- Output Power: Increased from the dispatched 210W to approximately 360W, which was the power demand of the local load, confirming successful autonomous power balancing in the island.
This test conclusively proved the solar inverter’s capability to maintain uninterrupted power supply during a grid outage.
Scenario 2: Grid Recovery and Reconnection. Starting from the SLI mode, the grid voltage was restored. The solar inverter initiated the pre-synchronization routine. Within six grid cycles, its output voltage was perfectly aligned with the grid in phase and amplitude. The STS was closed, and the control transitioned back to GC mode.
- The DC-link voltage remained stable during the entire process.
- The PV operating point smoothly transitioned back to the GC setpoint.
- The inverter output power immediately followed the $P_{ref}$ command of 210W.
- The total harmonic distortion (THD) of the inverter’s output voltage in GC mode was measured to be 2.88%, indicating high power quality.
This test validated the effectiveness of the pre-synchronization and bumpless reconnection strategy for the solar inverter.
Conclusion
This article has presented a comprehensive solution for enhancing the resilience and functionality of distributed solar inverters. By implementing a multi-modal control architecture, a standard grid-following solar inverter can be endowed with grid-forming capabilities, enabling it to operate as a storage-less islanded power source. The proposed seamless transfer strategy, based on state-initialization and active pre-synchronization, ensures smooth, transient-free transitions between grid-connected and islanded modes. Experimental results confirm that the strategy maintains voltage and frequency stability, manages power balance autonomously, and prevents harmful current surges during switching events.
The significance of this work lies in its potential to increase the penetration of PV systems without proportionally increasing reliance on costly energy storage for basic reliability. It empowers solar inverters to provide active grid support and maintain local power supply during disturbances, thereby improving power quality and reliability at the distribution level. This control paradigm represents a step forward in the evolution of solar inverters from simple power injection devices to intelligent, resilient, and grid-supportive assets in the future renewable-dominant power system.