As China’s energy transition advances alongside the ambitious “dual carbon” objectives, the installed capacity of distributed photovoltaic systems has been rising steadily. The active support control technology for the power grid has emerged as a critical research focus in recent years. For the distributed solar inverter operating in an extreme mode that loses support from the grid or energy storage, it becomes necessary to switch to an islanding mode without energy storage. In this state, a voltage-controlled strategy is adopted to independently support the voltage. Achieving a smooth transition between grid-connected operation and islanding operation without energy storage is of paramount importance for ensuring uninterrupted power supply to loads, particularly when the voltage synchronization signal is lost. In this study, I investigate multi-mode control strategies for the solar inverter without energy storage and propose a seamless transfer strategy. This strategy maintains stable operation of the solar inverter in both grid-connected and islanding modes without energy storage, and facilitates a smooth transition between these two modes, thereby ensuring power balance between the solar inverter and the load. Finally, experimental results validate the effectiveness of the proposed approach, ensuring the quality of power supply in distribution networks containing distributed photovoltaic systems across multiple operational modes.
Introduction and Motivation
Photovoltaic power generation, along with other renewable energy technologies such as wind power, offers significant advantages in terms of cleanliness, environmental friendliness, and sustainability. These technologies play an increasingly vital role in driving China’s energy transition toward green and low-carbon goals. With the introduction of the “carbon peak, carbon neutrality” targets and the vision of a new-type power system, the installed capacity of renewable energy sources like photovoltaic and wind power is expected to expand further, becoming an even more critical component of the future power system.
Unlike traditional fossil fuel-based power generation, the output power of photovoltaic systems is highly influenced by environmental factors. Consequently, power must typically be converted into stable alternating current through an inverter before being fed into the main grid and supplied to loads. As a key element in energy conversion, the structure and control strategy of the inverter are crucial for the safe and stable operation of the photovoltaic power generation system. Inverter control strategies can be broadly classified into voltage-controlled and current-controlled methods. Currently, most solar inverters operate with current-controlled strategies during grid connection. This control method relies on the main grid to provide voltage and frequency support. In extreme situations, such as grid faults leading to instability in bus voltage and frequency, solar inverters employing current-controlled technology will activate anti-islanding protection and cease operation, resulting in load disruption and reduced power quality for users.
To address this issue, a common approach involves equipping photovoltaic systems with energy storage. The bidirectional power capability of energy storage systems, which is highly flexible and controllable, allows the photovoltaic and storage system to form an island after a grid fault, providing uninterrupted power supply to local loads. However, energy storage capacity is limited and costly. When the storage is overcharged or deeply discharged, it may fail to provide a stable voltage synchronization signal, causing the photovoltaic system to shut down. Consequently, research has begun to explore voltage-controlled technologies for solar inverters without energy storage. In such scenarios, the solar inverter must independently support the bus voltage and frequency. This necessitates a switch from the grid-connected current-controlled mode to the islanding voltage-controlled mode. Therefore, this work focuses on transitioning the solar inverter from current-controlled to voltage-controlled operation during grid faults, enabling it to operate in an islanding mode without energy storage and ensuring uninterrupted power supply to loads.
In grid-connected mode, maximum power point tracking (MPPT) control is typically employed to maximize the utilization of solar energy by tracking the maximum power point of the photovoltaic array. However, due to variations in environmental factors, the output power of the photovoltaic system fluctuates in real-time, causing power fluctuations in the grid and acting as a disturbance source. Additionally, when the distribution network cannot fully absorb the power generated by the photovoltaic system, voltage violations at the point of common coupling can occur. In photovoltaic-storage islands, when the storage power and energy reach their limits, it becomes necessary to shed loads or reduce photovoltaic output. This involves a trade-off between economic efficiency and power supply quality to maintain the safe and stable operation of the grid. Such approaches have been explored to varying degrees in existing literature.
To mitigate power fluctuations caused by environmental changes, some studies have employed proportional-integral (PI) controllers to achieve closed-loop control of the solar inverter’s output power, enabling it to operate in a constant power output mode during grid connection. This approach allows the photovoltaic system to accept dispatch and management from higher-level control systems. Other studies have proposed MPPT strategies based on fuzzy control to enhance tracking speed and address issues of significant power fluctuations under varying irradiance levels. For islanded microgrids with multiple storage units and photovoltaic systems, adaptive power control methods based on AC bus voltage signals have been proposed to prevent overcharging of storage systems. Research on grid-connected and islanding switching has primarily focused on mitigating voltage and current transients during hard switching. For instance, some works have proposed seamless switching strategies based on virtual synchronous generator control, using the same control structure for both grid-connected and islanding modes to achieve smooth transitions. Other methods employ state followers to ensure continuity during the switching process, addressing the issue of current transients associated with hard switching. Nonlinear droop control strategies have also been introduced to adjust power output before transitioning from grid-connected to islanding operation, suppressing voltage and current surges. Pre-synchronization control schemes based on frequency perturbation have been proposed to alleviate overcurrent transients during switching. Furthermore, most existing switching strategies for photovoltaic systems consider scenarios with energy storage, i.e., switching between grid-connected and islanding modes in a photovoltaic-storage microgrid. In some cases, when the distribution network experiences a fault or when the storage output reaches its power limit and cannot maintain the PCC bus voltage within allowable limits, the photovoltaic-storage microgrid switches to islanding operation. Improved voltage regulators have been developed to achieve fast and seamless transitions between grid-connected and islanding modes, reducing current transients and bus voltage oscillations. Control structure improvements for storage inverters in photovoltaic-storage microgrids have also been proposed, sharing the current inner loop and compensating for the output current reference of the outer loop to mitigate transient impacts. However, in these switching strategies, the storage inverter’s control structure switches between voltage-controlled and current-controlled modes, while the photovoltaic system operates solely as a current-controlled source and does not actively participate in supporting the bus voltage. Moreover, these switching methods are primarily designed for microgrids with storage, with little discussion on islanding with load support for a solar inverter without energy storage and its associated mode transitions.
Building upon existing research on photovoltaic island microgrids, this work further considers the extreme scenario where no energy storage is available to provide a voltage synchronization signal. Therefore, I analyze the operational modes of the solar inverter without energy storage, study multi-mode control strategies, and propose a seamless transfer strategy for the solar inverter between grid-connected and islanding states. This strategy allows the solar inverter to switch from current-controlled grid-connected mode to voltage-controlled islanding mode after losing the voltage signal provided by the grid, thereby operating in an islanding mode without energy storage and providing uninterrupted power supply to loads. Finally, experimental verification confirms the effectiveness of the proposed control strategy, demonstrating its ability to maintain active and reactive power balance during the transition between grid-connected and islanding operations.
Multi-Mode Control of Solar Inverter without Energy Storage
The distributed solar inverter considered in this work primarily operates in four modes: grid-connected mode, islanding mode without energy storage, grid-connected transition mode, and islanding transition mode.
Operational Mode Analysis and Topology Selection
In grid-connected mode, the solar inverter employs a current-controlled strategy. Under this condition, the voltage and frequency of the AC bus are determined by the grid, and load fluctuations causing power imbalances between generation and consumption are compensated by the grid. Since the photovoltaic system needs to provide active power support to the grid, it operates in a constant power mode during grid connection. This means the solar inverter always tracks the power command from the higher-level system, switching to MPPT operation only when it cannot meet the command requirements.
When a grid fault occurs and the grid can no longer provide stable voltage and frequency support, the solar inverter should disconnect from the grid, enter the islanding transition mode, and switch to the islanding mode without energy storage for load support. In this islanding mode without energy storage, the solar inverter must employ a voltage-controlled strategy to support the bus voltage and balance load power demand. The solar inverter operates as a single unit in islanding mode, and constant voltage constant frequency (V/f) control can be employed to meet the voltage and frequency support requirements of the load. The key is real-time power balance between the source and load, i.e., regulating the photovoltaic output based on load power demand. When the grid’s power supply capability is restored, the solar inverter enters the grid-connected transition mode to resume operation as a current source.
Since the studied state requires flexible regulation of the port voltage of the solar inverter, a two-stage topology is adopted. This topology consists of a front-end Boost converter cascaded with a rear-end inverter, enabling decoupled control of the DC bus voltage and the output power of the photovoltaic array. Furthermore, due to the advantages of lower inductor current ripple, reduced switching stress, and lower switching losses offered by three-level inverters, this work selects a three-level Boost converter and a neutral point clamped (NPC) three-level inverter as the main circuit topology for the solar inverter. The solar inverter is connected to the main grid through a solid-state switch K1. When K1 is closed, the solar inverter operates in current-controlled grid-connected mode, supplying power to the load together with the grid. When the power quality at the AC bus is detected to be unsatisfactory for the load, K1 is opened, and with only the photovoltaic array as the power source, the solar inverter switches to islanding mode without energy storage to ensure uninterrupted load supply.
Grid-Connected Control Strategy
The control strategy for the solar inverter in this work fully utilizes the photovoltaic system’s regulation capability to support the bus voltage and frequency across multiple operational modes. Therefore, in grid-connected mode, the solar inverter does not use MPPT control for grid support. Instead, it employs a constant power tracking strategy, enabling the photovoltaic system to track power commands from the higher-level system with high performance, thereby facilitating stable grid voltage regulation.
In the rotating reference frame, the active and reactive power output of the solar inverter can be expressed as follows:
$$ P = \frac{3}{2} u_{md} i_{md} + \frac{3}{2} u_{mq} i_{mq} $$
$$ Q = \frac{3}{2} u_{mq} i_{md} – \frac{3}{2} u_{md} i_{mq} $$
where \( i_{md} \) and \( i_{mq} \) are the d-axis and q-axis components of the inverter output current \( i_m \) in the rotating reference frame, and \( u_{md} \) and \( u_{mq} \) are the d-axis and q-axis components of the inverter output voltage \( u_m \) in the rotating reference frame.
During grid-connected operation, the presence of a phase-locked loop (PLL) ensures that the grid voltage reference frame aligns with the inverter’s reference frame, allowing the q-axis component to be approximated as zero. Thus, the power equations can be simplified as follows:
$$ P = \frac{3}{2} u_{md} i_{md} $$
$$ Q = – \frac{3}{2} u_{md} i_{mq} $$
Applying these simplified equations to the inverter control enables tracking of the desired grid-connected power. However, this approach relies on power balance between the input and output terminals of the inverter, which manifests as the stability of the DC bus voltage between the Boost converter and the inverter.
Consequently, DC bus voltage stability is a prerequisite for stable grid-connected operation of the solar inverter. This work directly sets the DC bus voltage reference \( u_{dc\_ref} \) and uses the Boost converter to control the DC side voltage stability of the solar inverter. By adjusting the output power of the photovoltaic array based on changes in the DC voltage, the photovoltaic system ultimately tracks the power command from the higher-level system. When the output power of the photovoltaic system is limited by the maximum power point and cannot meet the higher-level command, the solar inverter reverts to operating near the maximum power point. Additionally, the Boost converter employs a voltage and current double closed-loop control to track the voltage command. Here, \( i_{L\_ref} \) is the current loop command value output by the voltage loop, used to control the inductor current \( i_L \) of the Boost converter. The rear-end inverter uses a current single closed-loop control technique, calculating the current loop command values \( i_{dq\_ref} \) based on the given power command values \( P_{ref} \) and \( Q_{ref} \).
The following table summarizes the key control parameters and strategies for the solar inverter in grid-connected mode:
| Parameter / Strategy | Description |
|---|---|
| Control Type | Current-controlled (PQ control) |
| Power Tracking | Constant Power Tracking (CPT) |
| DC Bus Voltage Control | Front-end Boost converter with voltage-current double loop |
| Inverter Control | Current single loop based on power command |
| Reference Frame | Synchronized with grid via PLL (q-axis ≈ 0) |
| Key Equations | P = 1.5 * u_md * i_md, Q = -1.5 * u_md * i_mq |
Islanding Control Strategy without Energy Storage
When a grid fault occurs and the grid can no longer provide stable voltage and frequency support, the solar inverter switches to islanding mode without energy storage. Unlike grid-connected mode, since external support for the AC bus is lost, the photovoltaic system must stabilize the voltage and frequency of the AC bus independently. Therefore, constant voltage and constant frequency control is adopted on the inverter side. Furthermore, when the solar inverter operates in islanding mode without energy storage, it must independently achieve power balance between generation and load. Specifically, when environmental factors cause changes in the output power of the photovoltaic array or when load demand changes, the photovoltaic array should rapidly adjust its operating point to maintain balance between the generation side output power and the load demand power.
Based on the two-stage solar inverter structure, power generated by the photovoltaic array is transmitted to the inverter side via the DC capacitor. When the photovoltaic array output power exceeds the load demand, the DC capacitor voltage rises. Conversely, when the output power is less than the load demand, the DC capacitor voltage falls. The power difference between the source and load sides is reflected as voltage fluctuations on the DC capacitor. Therefore, by stabilizing the DC bus voltage, source-load power balance can be achieved, determining the operating point of the front-end photovoltaic system and addressing power fluctuation issues.
The following table outlines the control parameters for the islanding mode without energy storage:
| Parameter / Strategy | Description |
|---|---|
| Control Type | Voltage-controlled (V/f control) |
| Voltage Reference | Constant amplitude (e.g., 70 V) |
| Frequency Reference | Constant frequency (e.g., 50 Hz) |
| Power Balance | Achieved via DC bus voltage regulation |
| Inverter Control | Voltage outer loop, current inner loop |
| Key Feature | No external energy storage source required |
Seamless Transfer Strategy for Solar Inverter without Energy Storage
In the previous section, I analyzed the control strategies for the solar inverter in both grid-connected and islanding modes without energy storage, achieving reliable operation in these two modes. In this section, I investigate the switching methods from grid-connected to islanding mode without energy storage and from islanding to grid-connected mode, further ensuring reliable operation of the distributed solar inverter across multiple modes.
Transition from Grid-Connected Mode to Islanding Mode without Energy Storage
When a grid fault is detected, the control structure of the solar inverter switches from grid-connected mode to islanding mode without energy storage. During this transition, the Boost converter continuously aims to stabilize the DC voltage, while the inverter control strategy switches from constant power control in grid-connected mode to constant voltage and constant frequency control in islanding mode without energy storage. To achieve a smooth transition from grid-connected mode to islanding mode without energy storage, a specific switching control strategy is adopted.
In grid-connected operation, the coordinate transformation reference phase \( \theta_{ref} \) is the grid-side phase \( \theta_{PLL} \) generated by the PLL. The current command \( i_{dq\_ref} \) corresponding to the power commands \( P_{ref} \) and \( Q_{ref} \) is calculated based on the simplified power equations. When a grid fault is detected and the system switches to islanding mode without energy storage, the current command \( i_{dq\_ref} \) is provided by the voltage outer loop, and the phase \( \theta_{ref} \) is generated by integrating the frequency command \( \omega_{ref} \). For the current inner loop control of the inverter, the control structure is essentially the same in both modes. The primary difference lies in the source of the inner loop command \( i_{dq\_ref} \), which can lead to overcurrent transients during switching if not handled properly. To avoid command jumps causing overcurrent transients, an integrator initialization method is adopted, allowing the integrator to seamlessly inherit the pre-switching values. When transitioning from grid-connected to islanding operation without energy storage, the command output by the voltage outer loop control inherits the grid-connected current command. The phase generator of the islanding control module without energy storage inherits the grid phase \( \theta_{PLL} \) from the PLL. This ensures that the current command and phase do not jump during the transition, thereby achieving a smooth transfer.
The following formula summarizes the phase generation during the transition:
$$ \theta_{ref} = \int \omega_{ref} \, dt \quad \text{(after transition, initial value set to } \theta_{PLL} \text{)} $$
Transition from Islanding Mode without Energy Storage to Grid-Connected Mode
Since the frequency and amplitude of the grid voltage fluctuate over time, the load voltage during islanding operation without energy storage may deviate from the grid voltage in terms of amplitude and frequency. To ensure successful grid connection, pre-synchronization control is required when transitioning from islanding mode without energy storage to grid-connected mode. This synchronization ensures that the load voltage matches the grid voltage, preventing overcurrent transients at the moment of grid connection due to phase mismatch.
In islanding mode without energy storage, the reference for voltage amplitude and phase is generated internally. When a grid-connection command is received, pre-synchronization begins. The grid-side voltage \( u_{dq\_PLL} \) is compared with the voltage command \( u_{mdq\_ref} \). The difference is integrated to obtain the voltage adjustment \( \Delta u_{dq} \), which is added to the command value \( u_{mdq}^{*} \) from islanding operation to generate the pre-synchronized voltage command \( u_{mdq\_ref} \). This allows the islanding voltage command to gradually adjust until it aligns with the grid voltage. Simultaneously, the grid-side phase \( \theta_{PLL} \) generated by the PLL is compared with the phase reference \( \theta_{ref} \). The phase difference \( \Delta \theta \) is processed by a PI controller to generate a frequency adjustment \( \Delta \omega \). When \( \theta_{PLL} > \theta_{ref} \), the compensation \( \Delta \omega > 0 \), increasing the angular frequency of the solar inverter so that \( \theta_{ref} \) gradually aligns with \( \theta_{PLL} \). When they are completely aligned, the PI controller output remains constant, and pre-synchronization is complete. Conversely, when \( \theta_{PLL} < \theta_{ref} \), the compensation \( \Delta \omega < 0 \), decreasing the angular frequency to achieve alignment.
Once the grid voltage and load voltage are in phase and matching in amplitude, pre-synchronization is complete. The coordinate transformation reference phase is switched to the grid phase from the PLL, and the solar inverter transitions to grid-connected mode.
The pre-synchronization control can be expressed by the following equations:
For voltage amplitude synchronization:
$$ u_{mdq\_ref} = u_{mdq}^{*} + \Delta u_{dq} = u_{mdq}^{*} + \int (u_{dq\_PLL} – u_{mdq\_ref}) \, dt $$
For phase synchronization:
$$ \theta_{ref} = \int (\omega_{ref} + \Delta \omega) \, dt \quad \text{where} \quad \Delta \omega = K_p \Delta \theta + K_i \int \Delta \theta \, dt $$
The following table summarizes the key differences between the two mode transition strategies:
| Transition Type | Key Control Action | Challenge Addressed |
|---|---|---|
| Grid-connected → Islanding (without energy storage) | Inherit current command and phase from pre-switching state | Prevent command jump and overcurrent |
| Islanding (without energy storage) → Grid-connected | Voltage and phase pre-synchronization using PI controllers | Align load voltage with grid voltage, prevent phase mismatch |
Experimental Validation
To validate the proposed seamless transfer strategy, I constructed a solar inverter experimental platform. The platform consists of a front-end Boost converter and a rear-end NPC three-level inverter, followed by an LCL filter, a local load, and a connection to the main grid via solid-state switches and air circuit breakers. A DC source in series with a resistor is used to emulate the external characteristics of a photovoltaic array. The controller is based on a DSP28377. The primary experimental parameters are summarized in the table below.
| Parameter | Value |
|---|---|
| Grid voltage (\( u_g \)) | 70 V |
| Islanding voltage command (\( u_{md\_ref} \)) | 70 V |
| Grid-connected power command (\( P_{ref} \)) | 210 W |
| Islanding frequency reference (\( f_{ref} \)) | 50 Hz |
| Load resistance (\( R_{load} \)) | 20 Ω |
| DC source voltage (\( u_{dc} \)) | 70 – 80 V |
| Switching frequency (\( f_k \)) | 10 kHz |
| Series resistance (\( R_{dc} \)) | 2.5 Ω |
To ensure load supply reliability, the solar inverter must rapidly transition from grid-connected to islanding mode without energy storage upon a grid fault, and resume grid-connected operation once the fault is cleared. During the experiment, upon detecting a grid fault, the solid-state switch connecting the solar inverter to the grid is opened, and the solar inverter switches from grid-connected mode to islanding mode without energy storage. When the grid recovers, the solar inverter first undergoes pre-synchronization. After the inverter output voltage aligns with the grid voltage, the solid-state switch closes, and the solar inverter transitions from islanding to grid-connected operation.
The experimental results for the transition from grid-connected mode to islanding mode without energy storage are presented here. The solar inverter initially operates in grid-connected mode, delivering 210 W of power. When an air circuit breaker is opened to simulate a grid fault, the inverter detects the fault, opens the solid-state switch, and simultaneously switches to islanding mode without energy storage. The photovoltaic port voltage \( u_{pv} \) changes from 74 V (during grid connection) to 70 V. The DC bus voltage \( u_{dc} \) experiences a slight dip before recovering to 200 V. The inverter output voltage \( u_m \) (which is also the load voltage) and the inverter output current \( i_m \) transition smoothly without any transients. The solar inverter output power tracks the load demand, increasing to 360 W. These results confirm that the proposed control strategy enables the solar inverter to seamlessly transition from grid-connected operation to islanding operation without energy storage after a grid fault.
For the reverse transition, the solar inverter initially operates in islanding mode without energy storage, with the grid voltage having been restored. The inverter output power is 360 W. Upon receiving the grid-connection command, pre-synchronization begins. After six power frequency cycles, the inverter output voltage \( u_m \) and the grid voltage \( u_g \) become synchronized. The solid-state switch closes, and the solar inverter transitions to grid-connected mode. During the transition, the DC bus voltage remains stable, the photovoltaic operating point transitions smoothly, and the solar inverter output power settles to 210 W, matching the grid-connected power command.
Furthermore, the total harmonic distortion (THD) of the solar inverter’s output voltage in grid-connected mode was measured. The THD for phase A was found to be 2.88%, indicating low harmonic content. These experimental results validate the effectiveness of the control strategy in achieving smooth transitions between grid-connected and islanding modes without energy storage for the solar inverter.
Conclusion
In this work, I addressed the scenario where a distributed solar inverter loses support from external power sources. I investigated multi-mode control strategies for the solar inverter without energy storage and proposed a seamless transfer strategy for the solar inverter between grid-connected and islanding operations without energy storage. After losing grid support, the solar inverter smoothly transitions from grid-connected mode to islanding mode without energy storage, thereby providing uninterrupted power supply to loads. Upon grid restoration, it smoothly transitions back to grid-connected mode. Compared to traditional switching strategies for solar inverters, the proposed control approach fully exploits the stable operating capability of the photovoltaic system below its maximum power point, reducing reliance on external storage sources or other power sources. This enables a smooth transition for the solar inverter between grid-connected and islanding states without significant voltage distortion or current transients during the switching process. Through theoretical analysis and experimental verification, I have demonstrated the power supply capability of the solar inverter under complex operating conditions, enhancing its active support capability for the grid. This expands the application scenarios for photovoltaic power generation and supports the deeper integration of distributed photovoltaic systems in distribution networks as the penetration of renewable energy continues to increase.
The key contributions and findings of this work can be summarized in the following table:
| Aspect | Finding / Contribution |
|---|---|
| Control Strategy | Multi-mode control (grid-connected PQ, islanding V/f without storage) |
| Switching Method | Seamless transfer using integrator initialization and pre-synchronization |
| Grid → Islanding | Smooth transition, no overcurrent; power increases from 210 W to 360 W |
| Islanding → Grid | Pre-synchronization completed in ~6 cycles; power returns to 210 W |
| Power Quality | Output voltage THD = 2.88% in grid-connected mode |
| Key Advantage | Eliminates need for energy storage in transient support; improves load reliability |
In summary, the proposed strategy ensures that the solar inverter can effectively support loads in a distribution network during extreme events without relying on energy storage. By leveraging the photovoltaic system’s inherent regulation capabilities through advanced control techniques, the solar inverter can maintain voltage and frequency stability, provide uninterrupted power supply, and facilitate a smooth return to grid-connected operation. This research contributes to the development of more resilient and flexible distributed generation systems, supporting the ongoing transition toward a higher share of renewable energy in the power grid.