The proliferation of distributed generation, particularly from photovoltaic (PV) sources, necessitates advanced power electronic interfaces. The grid-tied solar inverter is a critical component, responsible not only for injecting active power but also for maintaining or enhancing power quality as penetration levels rise. This paper explores a novel Two-Stage Quasi-Multifunctional Inverter (QMFI) topology designed to achieve high-efficiency power transfer while concurrently performing ancillary services such as reactive power compensation, harmonic filtering, and load balancing. Unlike conventional two-stage solar inverter systems where all power is processed twice, the proposed architecture enables a portion of the active power to flow through a single conversion stage, thereby improving overall system efficiency.
The core challenge addressed is the inherent efficiency penalty in traditional two-stage solar inverter systems. Typically, a front-end DC-DC converter (e.g., a Boost stage) performs Maximum Power Point Tracking (MPPT) and steps up the variable PV voltage. A subsequent DC-AC inverter then converts this stable DC voltage to grid-compatible AC power. While this decoupling offers control flexibility, every watt of power from the PV array passes through both conversion stages, incurring losses in each. The proposed QMFI topology ingeniously creates a direct power path, allowing a significant fraction of the generated power to bypass the DC-DC stage entirely, flowing directly from the PV array to the AC grid or load through the DC-AC stage.
Topological Architecture and Operational Principle
The proposed system, as shown in its functional block diagram, consists of two main power stages: a front-end DC-DC boost converter and a unique Dual-DC-Port (DDP) inverter. The PV array is connected to both stages. The boost converter manages the MPPT algorithm and regulates the high-voltage DC bus (UH). The innovative DDP inverter features not only the standard high-voltage DC bus input but also a second, low-voltage DC port (UL) connected directly to the PV array terminals. This low-voltage port, through the action of integrated bidirectional switches within each inverter leg, establishes the direct power flow path.
The key operational principle lies in the current splitting at the DC side. The total PV current (iPV) divides into two components: iL, which flows directly into the low-voltage port of the DDP inverter, and iH, which flows into the boost converter. Consequently, the corresponding power components are PL (single-stage power) and PH (two-stage power). The sum equals the total input power Pin. The DDP inverter synthesizes the AC output current (ix) which is the sum of the fundamental active current, and any compensating currents for non-active power (reactive, negative-sequence, harmonic) required by the local load. This enables the solar inverter to function as a multifunctional device, improving grid current quality while injecting active power.
Mathematical Modeling in Rotating Frame
To analyze and control the system effectively, a mathematical model in the synchronous rotating (dq) reference frame is essential. For the DDP inverter, the switching states for each phase leg (x = a, b, c) can be defined by an integer variable Stx representing the output voltage level relative to the negative DC rail:
$$S_{tx} = \begin{cases}
2, & \text{if connected to } U_H (S_{Hx}=1)\\
l, & \text{if connected to } U_L (S_{Lx1}, S_{Lx2}=1)\\
0, & \text{if connected to } 0 (S_{Zx}=1)
\end{cases}$$
where $l = U_L / E$ and $U_H = 2E$. The space vector diagram comprises 27 voltage vectors, categorized into large, medium, small (positive and negative), and zero vectors. To understand the power flow through the low-voltage port, a Current Switch Function (CSF) is introduced for each phase:
$$S_{ix} = \begin{cases}
1, & \text{if } S_{tx} = l\\
0, & \text{if } S_{tx} = 2 \text{ or } 0
\end{cases}$$
The current iL flowing into the low-voltage port during a switching period is the sum of contributions from the applied voltage vectors. For a reference vector synthesized using three nearest vectors (U0, U1, U2) with duty cycles d0, d1, d2, the current iL is given by:
$$i_L = \sum_{y=0,1,2} i_{Ly} = \sum_{y=0,1,2} \left\{ d_y [S_{iay} \ S_{iby} \ S_{icy}] \cdot \mathbf{I} \right\}$$
where $\mathbf{I} = [i_a, i_b, i_c]^T$ is the three-phase output current vector of the solar inverter. Crucially, only medium and small vectors contribute to iL, as large and zero vectors have CSFs that result in zero contribution in a three-wire system.
Transforming the three-phase currents into the dq-frame for harmonic order *h* (where *h=+1* for positive sequence fundamental, *h=-1* for negative sequence, etc.) simplifies analysis. In this frame, the current components ihd and ihq become constants. The expression for the *h*-th order component of iL becomes:
$$i_{Lh} = \sum \left\{ d_m [m_{hd} \ m_{hq}] \begin{bmatrix} i_{hd} \\ i_{hq} \end{bmatrix} + d_s[k \ 1-k] \begin{bmatrix} s_{hpd} & s_{hpq} \\ s_{hnd} & s_{hnq} \end{bmatrix} \begin{bmatrix} i_{hd} \\ i_{hq} \end{bmatrix} \right\}$$
Here, dm and ds are the duty cycles of the medium and small vectors, and the terms like mhd, shpd are influence factors derived from the CSFs and transformation matrices. This model is powerful because it allows for the analysis of how different voltage vector selections and different current components (active, reactive, harmonic) affect the single-stage power flow PL.
Modulation Strategy and Power Flow Analysis
The goal of the modulation strategy is to maximize the single-stage power flow PL to enhance efficiency. Based on the dq-frame model, the Average Value of Influence Factors (AVIF) over a fundamental line cycle can be calculated for different vector types and current components.
For the fundamental positive-sequence active current (i+d), analysis shows that the AVIF for medium vectors is zero, while the AVIF for positive small vectors is positive and for negative small vectors is negative. Therefore, to maximize iL (and hence PL), the modulation strategy should exclusively use positive small vectors for synthesis alongside the necessary medium and large vectors, while completely avoiding negative small vectors. This is a key differentiator for the control of this solar inverter topology.
The impact of compensating non-active currents on this beneficial power flow is analyzed using the same framework:
- Reactive Current Compensation (i+q): The AVIFs for both medium and positive small vectors related to i+q are zero. Therefore, compensating reactive power does not affect the single-stage power ratio PLr = PL/Pin.
- Negative-Sequence Current Compensation (i-d, i-q): All relevant AVIFs are zero. Compensating for load imbalance also does not perturb the pre-established direct power flow path.
- Harmonic Current Compensation (ihd, ihq): The effect varies with harmonic order. Low-order harmonics (e.g., 5th, 7th) may have non-zero AVIFs for the positive small vectors, potentially influencing PLr. However, higher-order harmonics have negligible impact. Even in the worst-case scenario with significant low-order harmonic compensation, a substantial portion of power (over 30% in the studied cases) still flows through the single-stage path, preserving the efficiency advantage.
The power distribution characteristic, defined by PLr, is primarily a function of the voltage ratio l = UL / E. A higher PV voltage UL leads to a greater share of power processed by the efficient single-stage path.
| Current Component | Medium Vector AVIF | Positive Small Vector AVIF | Effect on Single-Stage Power Flow (PL) |
|---|---|---|---|
| Fundamental Active (i+d) | ~0 | >0 | Deterministic, enables PL |
| Fundamental Reactive (i+q) | ~0 | ~0 | Negligible |
| Negative Sequence (i-d, i-q) | ~0 | ~0 | Negligible |
| 5th Harmonic | ~0 | ≠0 | Moderate Influence |
| 7th+ Harmonic | ~0 | ~0 | Negligible |
Control System Design
The control structure for the QMFI is implemented in the dq-domain. The front-end boost converter performs standard MPPT using an algorithm like Perturb & Observe, regulating its output to maintain a stable high-voltage DC bus UH.
The DDP inverter control is more involved. A Phase-Locked Loop (PLL) synchronizes with the grid. The load currents (iLa, iLb, iLc) are measured and transformed to the dq-frame. A Low-Pass Filter (LPF) extracts the DC components, which represent the fundamental positive-sequence load current. The non-active AC components (harmonic and negative-sequence) are obtained by subtraction. For power quality compensation, these AC components are added to the active current reference. The active current reference itself is generated by an outer DC-bus voltage control loop regulating UH; its output id* represents the total active power required from the PV array. A standard PI-based current controller tracks the final d-axis and q-axis reference currents.
An important nuance in the controller design for this solar inverter concerns the voltage loop. The plant model for the DC-bus voltage control differs from a conventional inverter because only a fraction (PHr = 1 – PLr) of the total power is processed through this bus. The open-loop transfer function Gv_open(s) includes this term:
$$G_{v\_open}(s) = G_v(s) \cdot \frac{G_{i\_open}(s)}{1+G_{i\_open}(s)} \cdot \frac{3v_d}{2V_H} \cdot (1 – P_{Lr}) \cdot \frac{1}{sC} \cdot H_v$$
Consequently, the proportional and integral gains (Kvp, Kvi) of the voltage PI controller Gv(s) should be adapted based on the operating point (i.e., the value of PLr) to maintain consistent bandwidth and stability margins across the PV voltage operating range.
Experimental Verification and Performance
A 3 kVA laboratory prototype was built to validate the proposed QMFI topology and its control strategies. Key parameters included a DC bus voltage UH = 700 V, a variable PV voltage UL ranging from 250 V to 600 V, switching frequencies of 50 kHz for the boost stage and 20 kHz for the DDP inverter, and appropriate LCL filter components.
The prototype successfully demonstrated all multifunctional capabilities. The MPPT algorithm effectively tracked the maximum power point, adjusting UL accordingly. The solar inverter seamlessly injected active power into the grid. Upon enabling compensation modes, the device accurately supplied reactive current, balanced unbalanced loads by injecting negative-sequence currents, and attenuated harmonic currents from non-linear loads, significantly improving the grid current quality. The dynamic response during step changes in reference power or during the activation of compensation functions was stable and rapid.
A critical set of experiments measured the system efficiency and compared it to a traditional two-stage solar inverter (comprising a boost converter and a three-level T-type inverter). The efficiency was evaluated under pure active power injection and under combined active power injection and compensation (where half the converter rating was used for compensation).
| Operating Mode | Proposed QMFI Efficiency | Traditional Two-Stage Inverter Efficiency | Efficiency Advantage |
|---|---|---|---|
| Pure Active Power | 97.1% | 95.8% | +1.3% |
| Active + Reactive Comp. | 96.9% | 95.5% | +1.4% |
| Active + Unbalance Comp. | 96.8% | 95.4% | +1.4% |
| Active + Harmonic Comp. | 96.7% | 95.3% | +1.4% |
The results consistently show a clear efficiency gain of over 1.3% for the proposed QMFI across its entire functional range. This gain is directly attributable to the single-stage power flow path, which processes a substantial portion of the total power with lower loss. The experiments confirm that the power quality compensation functions have minimal detrimental effect on this efficiency benefit, making the QMFI a highly attractive solution for modern PV systems.
Conclusion
This paper has presented a comprehensive analysis of a Two-Stage Quasi-Multifunctional solar inverter (QMFI). The topology introduces a dual-DC-port inverter structure that creates a direct, single-stage power flow path from the PV array to the grid, alongside the conventional two-stage path via a boost converter. A detailed mathematical model in the rotating dq-frame was developed, leading to the derivation of an optimized Space Vector Modulation strategy that maximizes the beneficial single-stage power flow.
The analysis conclusively shows that the compensation of non-active currents—reactive power, load unbalance, and most harmonics—has a negligible impact on the efficiency advantage offered by the direct power path. Only very low-order harmonic compensation introduces a minor influence, but the overall system efficiency remains superior to conventional two-stage solutions. Experimental results from a 3 kVA prototype validate the theoretical findings, demonstrating successful MPPT, high-quality grid current injection, and effective multifunctional compensation capabilities, all while achieving a measured efficiency increase of more than 1.3%. This work establishes the QMFI as a promising and efficient solar inverter topology for future distributed generation systems that are required to be both efficient and grid-supportive.