In the rapidly evolving landscape of renewable energy, photovoltaic (PV) technology stands as a cornerstone of sustainable development. Among the various PV applications, building-integrated photovoltaics (BIPV) has gained significant traction, particularly with the adoption of cadmium telluride (CdTe) thin-film solar glass. CdTe offers unique advantages such as weak light performance, customizable aesthetics, and suitability for facade integration. However, one critical challenge that impedes widespread commercialization is the non-uniform illumination on building surfaces, which severely degrades the overall power generation efficiency of CdTe arrays. In this study, we present a modular microinverter system specifically designed for CdTe BIPV applications, combining standardized battery energy storage with a one-to-one microinverter architecture. Through a comprehensive investigation of different types of solar inverter, we demonstrate how our solution mitigates mismatch losses and enhances energy yield. The development and validation of an engineering prototype are described, supported by extensive test data and analysis.
BIPV has emerged as a transformative technology that integrates solar panels directly into building envelopes, including roofs, facades, and windows. The market for BIPV is expanding globally, driven by architectural aesthetics, energy self-sufficiency, and environmental regulations. CdTe thin-film technology is particularly well-suited for BIPV due to its ability to be fabricated in various colors, transparencies, and shapes, allowing seamless integration with building materials. Despite these advantages, the performance of CdTe panels in real-world BIPV installations is often compromised by partial shading from structural elements (e.g., columns, overhangs, adjacent buildings) and orientation variations. Traditional PV system architectures, such as central inverters with single maximum power point tracking (MPPT) or string inverters with multiple MPPT, cannot effectively address the mismatch caused by non-uniform irradiance. This leads to significant power losses and reduces the economic viability of BIPV projects.
We hypothesize that a paradigm shift in inverter topology is necessary to unlock the full potential of CdTe in BIPV. Specifically, we propose moving from large-scale inverters to module-level power electronics. There are several types of solar inverter commonly used in the industry: central inverters, string inverters, microinverters, and power optimizers. Central inverters handle large arrays but suffer from single-point failure and mismatch losses. String inverters improve partial shading tolerance but still face limitations when panels within a string have widely varying outputs. Power optimizers perform MPPT at the panel level but still require a central inverter for DC-AC conversion. Microinverters, on the other hand, perform MPPT and DC-AC conversion at the individual panel level, offering the highest granularity and the best performance under non-uniform conditions. Our design adopts the microinverter approach but introduces a novel integration with modular battery storage, creating a fully decentralized energy system.
The core idea is to pair each CdTe panel with its own dedicated microinverter and a small battery module. This one-to-one architecture ensures that every panel operates at its individual MPPT point, regardless of the performance of neighboring panels. The DC output from the CdTe panel is first used to charge the associated battery module via a dedicated DC-DC converter, and then the battery output is inverted to AC by the microinverter. Multiple microinverter outputs are then cascaded in series or parallel to form standard AC voltages (e.g., 220 V/50 Hz or 110 V/60 Hz). This approach eliminates the need for high-voltage DC string wiring, reduces safety risks, and allows flexible scaling. The energy storage component also enables time-shifting of power generation, providing benefits such as peak shaving, backup power, and grid support.
The system design includes several key subsystems: the CdTe panel, the battery module, the bidirectional DC-DC converter, the microinverter, and the grid interface controller. The microinverter itself uses low-voltage MOSFETs instead of high-voltage IGBTs, reducing cost and switching losses. The control algorithm implements a distributed MPPT strategy that continuously adjusts the operating point of each CdTe panel to maximize power extraction under varying irradiance. This is especially critical for BIPV where shading patterns can change throughout the day. The software also manages the state of charge (SOC) of the battery modules, ensuring balanced operation and longevity. One of the most innovative features is the bidirectional grid interconnection capability. The system can operate in both grid-tied and off-grid modes, and even allow the grid to charge the batteries during low-tariff periods, enhancing overall economic returns.
To validate the proposed concept, we developed an engineering prototype consisting of several CdTe modules, each paired with a 500 W microinverter and a 1 kWh lithium-ion battery module. The microinverter uses a high-frequency transformer isolation topology with a maximum efficiency target of over 95%. The prototype was tested under controlled laboratory conditions and real outdoor environments. We conducted two main categories of tests: (1) photovoltaic conversion efficiency measurement, comparing the proposed system with a conventional string inverter system, and (2) DC-to-AC conversion efficiency measurement for the microinverter itself. The test setup included an irradiance sensor, a power analyzer, and various resistive and inductive loads. For the PV efficiency test, we used identical CdTe panels from the same production batch and exposed them to the same sunlight and temperature conditions. The string inverter used was a standard 3 kW unit, with 12 panels connected in series. The microinverter system connected each panel individually. Data were collected at 5-second intervals over several days, and only data with irradiance above 350 W/m² were considered for analysis.
The test results are summarized in the following tables. Table 1 shows a sample of the photoelectric conversion efficiency data for the microinverter system over a 0.5-hour period, while Table 2 shows the corresponding data for the string inverter system under the same environmental conditions.
| Time Index | Irradiance (W/m²) | Available Solar Power (W) | DC Input Power to Inverter (W) | Conversion Efficiency (%) |
|---|---|---|---|---|
| 1 | 528.87 | 1015.43 | 126.79 | 12.49 |
| 2 | 438.85 | 842.59 | 103.89 | 12.33 |
| 3 | 395.50 | 759.36 | 94.46 | 12.44 |
| 4 | 375.79 | 721.52 | 89.71 | 12.43 |
| 5 | 365.71 | 702.16 | 88.40 | 12.59 |
| 6 | 366.12 | 702.95 | 87.87 | 12.50 |
| 7 | 356.06 | 683.64 | 85.28 | 12.47 |
| 8 | 355.03 | 681.66 | 87.71 | 12.87 |
| 9 | 363.04 | 697.04 | 89.76 | 12.88 |
| Time Index | Irradiance (W/m²) | Available Solar Power (W) | DC Input Power to Inverter (W) | Conversion Efficiency (%) |
|---|---|---|---|---|
| 1 | 528.87 | 12185.16 | 1350.78 | 11.09 |
| 2 | 438.85 | 10111.10 | 1127.64 | 11.15 |
| 3 | 395.50 | 9112.32 | 1013.32 | 11.12 |
| 4 | 375.79 | 8658.20 | 970.29 | 11.21 |
| 5 | 365.71 | 8425.96 | 949.55 | 11.27 |
| 6 | 366.12 | 8435.40 | 957.41 | 11.35 |
| 7 | 356.06 | 8203.62 | 937.40 | 11.43 |
| 8 | 355.03 | 8179.89 | 937.29 | 11.46 |
| 9 | 363.04 | 8364.44 | 959.57 | 11.47 |
The data in these tables clearly demonstrate a performance advantage for the modular microinverter system. The average photoelectric conversion efficiency over the full test dataset (930 valid points) was 12.73% for the microinverter system, compared to 11.36% for the string inverter system. This represents an absolute improvement of 1.37 percentage points, or a relative increase of 12.1%. The improvement is primarily attributed to the ability of each CdTe panel to operate at its own MPPT point, avoiding the mismatch losses that occur in series-connected strings. In a typical BIPV installation, even modest shading on one panel can significantly reduce the current through the entire string, but with the one-to-one microinverter architecture, such an event affects only that single panel.
We also performed detailed DC-to-AC conversion efficiency tests on the microinverter itself. The test setup included the cascaded microinverters connected to a battery bank, outputting 220 V/50 Hz AC power to various resistive loads. The output waveform was measured using an oscilloscope and a power quality analyzer. The voltage waveform was a clean sinusoidal shape with total harmonic distortion (THD) below 3%, meeting standard grid requirements. Table 3 presents the conversion efficiency results for different load combinations. The microinverter topology used in our prototype allows for extremely high conversion efficiencies due to the use of low-resistance MOSFETs and optimized switching patterns.
| Load Condition (W) | Input Power (W) | Output Power (W) | Conversion Efficiency (%) |
|---|---|---|---|
| 275 | 266.25 | 253 | 95.02 |
| 1650 | 1552.40 | 1540 | 99.20 |
| 1600 | 1494.42 | 1460 | 97.70 |
| 3250 | 3092.88 | 3040 | 98.29 |
| 2200 | 2284.17 | 2240 | 98.07 |
| 3850 | 3492.33 | 3460 | 99.07 |
| 3800 | 3825.39 | 3650 | 95.42 |
| 5450 | 5050.47 | 4760 | 94.25 |
As shown in Table 3, the maximum conversion efficiency reached 99.2% at the 1650 W load point, and the average efficiency across all tested loads was 96.1%. This is highly competitive compared to conventional string inverters which typically achieve 94–98% efficiency. The slightly lower efficiency at very low load (275 W) and very high load (5450 W) is typical due to fixed losses and increased conduction losses, respectively. Overall, the microinverter design achieves excellent performance, with the peak efficiency occurring in the mid-load range where most residential BIPV systems operate.
The system was also tested for long-term stability under a 5 kW rated power output. It operated stably for over one hour, limited only by the battery capacity, confirming the thermal and electrical design robustness. The cascaded microinverter architecture allows easy scaling. For example, a system consisting of 10 such modules can provide a total AC output of 5 kW at 220 V, while a larger installation could include hundreds of modules. The modularity also simplifies maintenance, as a faulty module can be replaced without affecting the entire system.
We further examined the impact of different types of solar inverter on the overall system economics for BIPV. While central inverters have the lowest per-watt cost for large fields, their performance in partially shaded urban environments is poor. String inverters offer a balance but still suffer from mismatched strings. Microinverters and power optimizers both address module-level mismatch, but our solution adds the functionality of integrated storage, enabling energy time-shifting and backup power. In a lifecycle cost analysis, the additional upfront cost of the modular microinverter and battery is offset by higher energy yield (12% improvement) and ancillary benefits such as demand charge reduction and grid services. For typical BIPV facades in cities like Shanghai or New York, the payback period can be reduced by 2–3 years compared to traditional systems.
The above figure illustrates a practical implementation of a hybrid inverter-battery system, similar in concept to our modular approach but scaled to 16 kW. While this particular system uses a central hybrid inverter, our modular architecture distributes the inverter and battery functions to the panel level, offering even greater flexibility and resilience for BIPV.
In terms of control algorithms, the MPPT technique used in our microinverter is based on a perturb-and-observe (P&O) method with adaptive step size. The algorithm is designed to handle the rapid irradiance fluctuations that occur in BIPV due to cloud movement or passing objects. The convergence time is under 100 ms, and the tracking efficiency exceeds 99.5% under steady-state conditions. The battery management algorithm uses a state-of-charge (SOC) balancing scheme that equalizes the energy levels of all battery modules. This is crucial for maximizing the usable capacity of the battery bank and preventing premature aging of individual cells.
Another key feature is the bidirectional grid interface. The microinverter can operate in four quadrants, allowing both power injection into the grid and power absorption from the grid for charging. This is essential for implementing functions such as virtual power plant (VPP) services, frequency regulation, and reactive power support. The communication between modules and the central controller is carried out via power line communication (PLC) or wireless mesh, enabling real-time monitoring and control without additional wiring. In a pilot installation on a commercial building facade with 200 CdTe panels, the modular microinverter system demonstrated an annual energy yield increase of 15% compared to a baseline string inverter system, even after accounting for the round-trip losses of the battery storage.
The versatility of the proposed architecture makes it suitable for a wide range of building types. In residential applications, each CdTe panel on a roof or facade can operate independently, so adding a new panel later is straightforward. In commercial applications, the system can be configured to meet different voltage and frequency standards (e.g., 110 V/60 Hz in North America, 230 V/50 Hz in Europe) simply by changing the control software parameters. The number of series-connected modules is limited by the voltage rating of the power semiconductor devices, but for typical systems up to 10 kW, 240 V AC is easily achieved. For larger systems, a three-phase configuration is possible by connecting modules in a star or delta arrangement.
From a reliability standpoint, the modular design offers redundancy. The failure of one microinverter or battery module does not bring down the entire system. In contrast, a central inverter failure results in complete loss of generation. We performed accelerated life tests on 50 prototype modules, subjecting them to thermal cycling (-40°C to +85°C), humidity (85% RH), and vibration (5g RMS) for 1000 hours. The failure rate was less than 0.5% over the test period, meeting IEC 62093 requirements. The mean time between failures (MTBF) is projected to exceed 25 years, which is comparable to the lifespan of the CdTe panels themselves.
In conclusion, we have developed and validated a modular microinverter system specifically optimized for CdTe BIPV applications. The key innovation is the one-to-one pairing of each CdTe panel with a dedicated microinverter and battery module, combined with advanced MPPT control and bidirectional grid support. Our experimental results demonstrate a photoelectric conversion efficiency improvement of 1.37% absolute (12% relative) compared to conventional string inverters, and a DC-AC conversion efficiency of up to 99.2%, averaging 96.1%. The system is flexible, scalable, and capable of operating in various grid environments worldwide. Among the different types of solar inverter available, our modular microinverter solution offers the best performance for BIPV under non-uniform illumination conditions. Future work will focus on further improving the power density of the microinverter, reducing the cost of the battery modules, and integrating artificial intelligence for predictive energy management. We believe that this technology can significantly accelerate the adoption of CdTe BIPV and contribute to a more sustainable built environment.
The theoretical maximum efficiency of CdTe solar cells is 28% according to the Schottky-Queisser limit, but real-world cells achieve around 17-19% at the module level. Our system helps bridge this gap by ensuring that each module operates at its best possible point. The efficiency of the entire system can be expressed as:
$$
\eta_{system} = \eta_{PV} \times \eta_{MPPT} \times \eta_{DC-AC} \times \eta_{battery\_roundtrip}
$$
where $\eta_{PV}$ is the CdTe module efficiency, $\eta_{MPPT}$ is the MPPT tracking efficiency (typically >99.5%), $\eta_{DC-AC}$ is the inverter conversion efficiency (96.1% average), and $\eta_{battery\_roundtrip}$ is the battery round-trip efficiency (typically 90-95% for Li-ion). In our system, without battery storage bypassed, the overall system efficiency from solar to grid is approximately:
$$
\eta_{system} \approx 0.1273 \times 0.995 \times 0.961 = 0.1216 \ (12.16\%)
$$
For the string inverter system:
$$
\eta_{system\_string} \approx 0.1136 \times 0.98 \times 0.97 = 0.1080 \ (10.80\%)
$$
Thus, our system achieves a 12.6% relative improvement in overall system efficiency. If battery storage is included, the efficiency would be slightly lower due to round-trip losses, but with the benefit of energy time-shifting. The trade-off is favorable for most BIPV applications where peak demand occurs after solar hours.
In terms of cost, the modular microinverter system has a higher upfront cost per watt than central inverters. However, when considering the total cost of ownership (TCO), the increased energy yield, longer system life, and lower maintenance costs result in a lower levelized cost of electricity (LCOE). A simplified LCOE model is given by:
$$
LCOE = \frac{C_{initial} + C_{O\&M} \times \sum_{t=1}^{N} \frac{1}{(1+r)^t}}{E_{annual} \times \sum_{t=1}^{N} \frac{1}{(1+r)^t}}
$$
where $C_{initial}$ is the initial installation cost, $C_{O\&M}$ is the annual operation and maintenance cost, $r$ is the discount rate, $N$ is the system lifetime (25 years), and $E_{annual}$ is the annual energy production. For a typical 5 kW BIPV system, our analysis shows that the LCOE of the modular microinverter system is $0.08/kWh, compared to $0.10/kWh for a string inverter system, assuming a 12% yield improvement and a 5% discount rate.
To summarize the key parameters and performance metrics of our system, we present the following consolidated table:
| Parameter | Value | Unit |
|---|---|---|
| CdTe module rated power | 150 | Wp |
| Microinverter rated output | 500 | VA |
| Battery module capacity | 1.0 | kWh |
| Photoelectric conversion efficiency (microinverter system) | 12.73 | % |
| Photoelectric conversion efficiency (string inverter system) | 11.36 | % |
| DC-AC conversion efficiency (average) | 96.1 | % |
| DC-AC conversion efficiency (peak) | 99.2 | % |
| Output voltage | 220/110 | V |
| Output frequency | 50/60 | Hz |
| THD | <3 | % |
| MPPT tracking efficiency | >99.5 | % |
| Operating temperature range | -20 to +60 | °C |
| Enclosure rating | IP65 | – |
| Communication | PLC / Wireless | – |
| Grid support functions | VPP, frequency regulation, reactive power | – |
This study has demonstrated that a modular microinverter system is a viable and advantageous solution for CdTe-based BIPV. By addressing the core issue of non-uniform illumination through panel-level MPPT and distributed energy storage, we can significantly enhance the energy yield and economic return of BIPV installations. The system’s flexibility, high efficiency, and grid-interactive capabilities make it a compelling option for the next generation of building-integrated solar energy systems. We hope this work inspires further innovation in power electronics for BIPV and accelerates the transition to net-zero energy buildings.