We present a comprehensive discussion of the functional building blocks that enable photovoltaic (PV) inverters, also known as solar inverters, to operate effectively as distributed generators (DGs) within a smart grid environment. The rapid growth of PV installations, especially in low voltage distribution networks, has introduced new technical requirements for grid interconnection. Solar inverters, which are the essential interface between PV arrays and the electrical grid, must therefore incorporate a set of fundamental and advanced functions. In our work, we have developed and validated several core technologies for these building blocks, including a space-vector pulse-width-modulation (SV-PWM) based current controller, a golden-section-search (GSS) based maximum power point tracking (MPPT) strategy, and an accelerated automatic phase shift (AAPS) method for islanding detection. In addition, we have integrated communication and system control functions into solar inverters to support operation within virtual power plants and to allow PV systems to offer value-added services to power systems. This paper describes these building blocks in detail, presents the underlying mathematical formulations, and highlights their practical significance through experimental and field verification.
The motivation for enhancing solar inverter functionality stems from the fact that PV systems are no longer viewed as passive negative loads. With increasing penetration levels, PV DGs must actively support grid stability and power quality. The traditional building blocks of a solar inverter include power conversion, grid interconnection, and maximum power extraction. However, the smart grid context adds new layers of complexity, such as remote monitoring, dispatchability, and coordinated reactive power support. Therefore, a modern solar inverter should be built with a modular architecture that includes power stage hardware, control algorithms, protection mechanisms, and communication interfaces. We have organized these into the following building blocks: power conversion, grid interconnection, maximum power extraction, islanding detection, communication, and system-level control.
In the remainder of this article, we explain each building block in detail, starting with the current controller, which directly determines the output power quality of solar inverters.
1. SV-PWM Based Current Controller
Grid-connected solar inverters must deliver sinusoidal currents into the grid with low total harmonic distortion (THD), even when the grid voltage is distorted. We have developed a high-performance current controller based on space-vector pulse-width-modulation (SV-PWM) and a predictive control strategy in the synchronous d-q reference frame. This controller is designed to provide fast dynamic response and robust steady-state performance. The schematic structure of the control system consists of a predictor, a current error compensation block, and an SV-PWM modulator. The grid phase voltages and the grid angle are detected and used for feed-forward compensation of grid harmonics.
The predictive current controller is based on a deadbeat control concept. In each PWM cycle, the required inverter output voltages are calculated such that the actual currents will equal the reference currents at the end of the next cycle. The mathematical formulation in the d-q reference frame is given by the following equations:
\[
\begin{aligned}
u_d^*(n) &= L \frac{i_d^*(n) – i_d(n)}{T_s} – \omega L i_q(n) + u_{gd}^{av}(n+1) \\
u_q^*(n) &= L \frac{i_q^*(n) – i_q(n)}{T_s} + \omega L i_d(n) + u_{gq}^{av}(n+1)
\end{aligned}
\]
where \(u_d^*(n)\) and \(u_q^*(n)\) are the required d-axis and q-axis output voltages for the SV-PWM modulator at the \(n\)-th sampling instant; \(i_d^*(n)\) and \(i_q^*(n)\) are the reference currents; \(i_d(n)\) and \(i_q(n)\) are the measured actual currents; \(L\) is the filter inductance; \(\omega\) is the grid fundamental angular frequency; \(T_s\) is the PWM period; and \(u_{gd}^{av}(n+1)\) and \(u_{gq}^{av}(n+1)\) are the average grid voltages in the next PWM cycle, which are updated at a high sampling rate to include harmonic voltages. This feed-forward compensation of grid harmonics is a key feature of our controller, as it allows the output currents to remain sinusoidal even when the grid voltage itself contains significant harmonic distortion.
The SV-PWM modulator generates switching signals with a fixed switching frequency, which yields a well-defined harmonic spectrum and simplifies the design of the output filter. Compared with conventional proportional-integral (PI) controllers with sinusoidal pulse-width modulation, the SV-PWM based predictive controller offers advantages such as faster transient response, lower current THD, and better utilization of the dc-link voltage. In our experimental tests, under grid voltage harmonics up to 2.6% THD, the inverter output current remained highly sinusoidal, with negligible harmonic distortion. The current vector locus diagrams under step changes in the current reference demonstrate excellent dynamic behavior, with a settling time of only a few PWM cycles.
| Parameter | Value |
|---|---|
| Current THD under distorted grid (2.6% voltage THD) | < 1.5% |
| Dynamic response time | < 1 ms |
| Switching frequency | 10 kHz |
| Grid voltage feed-forward | d-q harmonic compensation |
2. GSS-Based Maximum Power Point Tracking
Maximum power point tracking (MPPT) is an essential function for solar inverters because the power-voltage characteristic of a PV array is nonlinear and the maximum power point (MPP) varies with irradiance and temperature. Many MPPT algorithms have been proposed in the literature, such as perturb-and-observe (P&O) and incremental conductance (INC). While these methods are popular due to their simplicity, they suffer from a trade-off between tracking speed and steady-state oscillations. To overcome these limitations, we have developed a multi-stage MPPT strategy that combines the golden section search (GSS) method with P&O and INC. The GSS method is a well-known optimization technique for finding the extremum of a unimodal function within an interval. In the context of PV power curves, the power as a function of voltage is unimodal, so GSS can be applied to locate the MPP rapidly and accurately.
The golden section search algorithm works by repeatedly narrowing the search interval. At each iteration, two interior points are selected according to the golden ratio \(\varphi\):
\[
\varphi = \frac{1+\sqrt{5}}{2} \approx 1.61803398
\]
If the search interval is \([a, b]\), we choose two points \(x_3\) and \(x_4\) such that:
\[
\frac{b-x_4}{x_4-a} = \varphi, \quad \frac{x_3-a}{b-x_3} = \varphi
\]
Then we evaluate the power at these two points. If the power at \(x_3\) is greater than the power at \(x_4\), the maximum must lie in the interval \([a, x_4]\); otherwise, it lies in \([x_3, b]\). This process is repeated until the interval width is smaller than a specified tolerance. The GSS method guarantees convergence with fewer function evaluations than many other methods.
In our multi-stage MPPT algorithm, the GSS method is used during the initial tracking or after large disturbances to quickly converge to the vicinity of the MPP. Once the system is close to the MPP, a small-step P&O or INC method is used for fine tracking. This hybrid approach combines the fast convergence of GSS with the precise steady-state behavior of P&O/INC. The flowchart of our GSS-MPPT algorithm was implemented in a digital signal processor and applied to the boost chopper stage of solar inverters. The measured MPPT process under fixed irradiance and temperature conditions is shown in the following experimental observations: the algorithm converged to the MPP within approximately 200 ms and maintained a very small steady-state power ripple. Field tests under varying irradiance also confirmed excellent tracking performance, with an average tracking efficiency above 99%.
| Method | Convergence speed | Steady-state oscillation | Complexity | Tracking efficiency |
|---|---|---|---|---|
| P&O | Moderate | High | Low | 95-98% |
| INC | Moderate | Medium | Medium | 96-99% |
| GSS-based (proposed) | Fast | Low | Medium | >99% |
The GSS-based MPPT has been integrated into commercial solar inverters and has proven robust in real-world operating conditions. Its ability to quickly converge under rapidly changing irradiance makes it especially useful in distributed PV systems where partial shading and cloud transients are common.
3. AAPS-Based Islanding Detection
Islanding detection is a critical protection function for all distributed generators, including solar inverters. When the utility supply is disconnected, the inverter must detect the island formation and cease energization within a specified time to prevent safety hazards and damage to equipment. Grid interconnection standards, such as IEEE 1547, UL 1741, and CSA C22.2 No. 107.1-01, require robust anti-islanding protection. Many islanding detection methods have been developed, and they can be broadly classified into remote methods and local methods. Remote methods, such as power line carrier and SCADA, are reliable but expensive and require coordination between the DG and the utility. Local methods are further divided into passive and active techniques. Passive methods monitor system parameters and detect abnormal changes, but they suffer from large non-detection zones when the DG power closely matches the local load. Active methods intentionally inject small disturbances into the inverter output and monitor the system response. Examples include active frequency drift (AFD), slide-mode frequency shift (SMS), and automatic phase shift (APS).
We have developed an improved islanding detection method based on accelerated automatic phase shift (AAPS). The fundamental principle of the AAPS method is to introduce a phase shift between the inverter output current and voltage that accelerates when an islanding condition is suspected. In a grid-connected mode, the phase shift has minimal effect because the grid imposes its voltage and frequency. However, if the grid is disconnected, the phase shift causes a cumulative frequency or period drift, which eventually trips the over/under frequency protection.
The basic phase shift of the AAPS method at the \((k+1)\)-th voltage cycle is given by:
\[
\theta_{\mathrm{AAPSB}}(k+1) = \pi \frac{T_{\mathrm{avg}} – T_v(k)}{T_v(k)}
\]
where \(T_v(k)\) is the measured period of the \(k\)-th voltage cycle, and \(T_{\mathrm{avg}}\) is the average period of the previous \(N\) cycles:
\[
T_{\mathrm{avg}} = \frac{1}{N} \sum_{i=k-N}^{k-1} T_v(i)
\]
During the next \(N\) voltage cycles after the \(k\)-th cycle, the algorithm evaluates the correlation between the applied phase shift and its impact on the voltage period. An additional phase shift is introduced based on a sign function of the period deviation:
\[
\theta_0(k+1) = \theta_0(k) + \Delta\theta \cdot \operatorname{sgn}(\Delta T)
\]
\[
\theta_{\mathrm{AAPS}}(k+1) = \theta_{\mathrm{AAPSB}}(k+1) + \theta_0(k+1)
\]
where \(\Delta\theta\) is a small constant phase increment, and \(\operatorname{sgn}(\Delta T)\) is the sign of \((T_{\mathrm{avg}} – T_v(j))\). If an island persists, the added phase shift will cause the period to change further in the same direction. The algorithm calculates the probability of cause and effect (PCE) by checking whether the following logical relationships hold:
\[
\theta_{\mathrm{AAPS}}(j) > 0 \implies \Delta T(j) = T_{\mathrm{avg}} – T_v(j) > 0
\]
\[
\theta_{\mathrm{AAPS}}(j) < 0 \implies \Delta T(j) = T_{\mathrm{avg}} – T_v(j) < 0
\]
If, after \(N\) voltage cycles, the PCE is greater than 0.6, an additional phase shift is applied and \(T_{\mathrm{avg}}\) is held constant. During periods with extra phase adjustment, the algorithm monitors the average of the previous \(N/2\) periods. If two consecutive averages show only a small difference, or if the PCE falls below 0.6, the additional phase shift is disabled, and the average period computation resumes. This adaptive behavior ensures that the inverter remains stable during grid-connected operation while providing accelerated detection under islanding conditions.
We have compared the AAPS method with the conventional APS method through simulation and experiments. In a test scenario where an island was formed at 0.5 s, the APS method failed to detect islanding because the frequency remained within the allowable operating window. In contrast, the AAPS method successfully detected the island and tripped the inverter quickly. Figure 2 (in our original publication) showed the grid voltage and current waveforms during a UL/CSA certification test, where the inverter shut down within 15 cycles (0.25 s) of island formation, well below the 2 s requirement.
| Feature | APS | AAPS |
|---|---|---|
| Phase shift adaptation | Constant feedback | Accelerated with PCE |
| Detection speed | Slow or fail under matched load | Fast and reliable |
| Non-detection zone | Relatively large | Small |
| Grid-connected disturbance | Low | Very low |
The developed AAPS method has been embedded in grid-connected solar inverters and wind inverters. It provides robust islanding protection in compliance with international standards while minimizing perturbs under normal grid-connected operation.
4. Communication and System Control Functions
As smart grid technologies advance, solar inverters are expected to participate actively in power system operation. In addition to the fundamental building blocks described above, modern solar inverters must include communication and system-level control functions. These advanced functions enable remote monitoring, aggregated dispatch, frequency regulation, reactive power support, and harmonic cancellation at the point of common coupling. We have developed a communication module that connects the inverter’s DSP controller to a virtual power plant environment. The module consists of two microcontrollers: one acquires measurement data from the inverter DSP, and the other displays the data on an LCD and transmits it to the internet via TCP/IP protocols. The data sent to the remote terminal include output voltage, current, active power, reactive power, energy production, power factor, fault status, and converter operation status. This information is essential for centralized energy management systems to coordinate multiple DGs and dispatch their outputs efficiently.
The communication module supports standard Ethernet and RS232 interfaces, making it compatible with a wide range of grid monitoring systems. Figure 3 illustrated the block diagram of the communication module, where the first microcontroller (1st uC) reads data from the DSP through a data bus, and the second microcontroller (2nd uC) handles the Ethernet communication and LCD display. The module also supports E2PROM for configuration storage and an interface for panel buttons and LEDs. In our field demonstration, the inverter operating data were transmitted to a remote terminal using TCP/IP, and the remote graphical display showed real-time PV output power variations over time, confirming the feasibility of remote monitoring and control.
| Function | Description |
|---|---|
| Real-time monitoring | Remote observation of voltage, current, power, and status |
| Aggregated power dispatch | Virtual power plant controls multiple solar inverters |
| Reactive power control | Voltage support through reactive injection or absorption |
| Frequency regulation | Active power curtailment or boost based on grid frequency |
| Harmonic cancellation | Injection of harmonic currents to reduce distortion at PCC |
With the integration of these advanced functions, solar inverters become intelligent assets in the smart grid. A virtual power plant can aggregate numerous PV systems, energy storage units, and controllable loads to provide services similar to conventional power plants. For example, based on solar power production forecasts, the virtual power plant can schedule aggregated PV output to participate in energy markets. During grid voltage deviations, the reactive power capability of solar inverters can be dispatched to support voltage. In the same manner, active power curtailment can be used for frequency regulation when over-frequency events occur. The communication module is therefore a critical enabler for these system-level functionalities.
5. Summary of Building Blocks
We summarize the complete set of building blocks for solar inverters in the following table. These blocks work together to ensure safe and efficient operation of PV systems in the smart grid:
| Building block | Key technology | Benefit |
|---|---|---|
| Power conversion | Three-phase inverter with SV-PWM | High efficiency, low harmonics |
| Maximum power extraction | GSS-based hybrid MPPT | Fast tracking, high energy yield |
| Grid interconnection | LCL filter, synchronization | Meets IEEE/UL/CSA standards |
| Islanding protection | AAPS active method | Fast detection, small NDZ |
| Communication | TCP/IP module | Remote monitoring and control |
| System control | Virtual power plant interface | Ancillary services, market participation |
All the technologies we have developed have been implemented and proven in commercial solar and wind inverters that are deployed in distributed generation markets across China and North America. Our experience indicates that a well-designed inverter should not only meet basic grid interconnection requirements but also integrate intelligent functions that add economic value. The ongoing evolution of smart grids will continue to demand more advanced capabilities from solar inverters, such as machine learning-based fault detection, grid-forming operation, and cybersecurity enhancements. These future building blocks will further strengthen the role of solar inverters as indispensable elements of the modern power system.
6. Conclusion
In this paper, we have presented the essential building blocks of distributed solar inverters for smart grid applications. The SV-PWM based current controller provides high-quality current regulation with low harmonic distortion even under distorted grid voltage conditions. The GSS-based MPPT algorithm ensures fast and accurate maximum power extraction from PV arrays under varying environmental conditions. The AAPS islanding detection method offers reliable anti-islanding protection with a negligible non-detection zone. In addition, communication and system control functions enable solar inverters to participate in virtual power plant operations and provide ancillary services to the grid. These building blocks together form a complete and proven solution for grid-connected PV systems. As the role of distributed energy resources continues to expand, the importance of multifunctional solar inverters in smart grids will only increase. We believe that continued innovation in inverter control and communication will unlock new opportunities for renewable energy integration.