Search

The metallization of silicon solar cells is often optimized for their performance under standard test conditions (STC) at 1000 W/m2 of AM1.5G illumination and 25 °C cell temperature. However, solar modules in field installations, experience a range of illumination intensities and distributions, spectral conditions and operating temperatures, depending on location and type of installation. In this work, we aim to answer the question of whether significant energy yield increases can be realized with solar cells that are optimized under average local conditions. We systematically optimize the metallization of monocrystalline PERC solar cells for a range of irradiance and temperature values. We find the optimum number of fingers to vary between 75 to 175 and the number of busbars between four to eight for 800 μm wide planar standard ribbons. We assessed the impact of these cell designs on energy yield of modules in a single-axis tracking scenario in Alice Springs, Australia, as well as in a building-integrated facade in Cabauw, Netherlands. A module with cells optimized for 600 W/m2 and 30°C, which is the average plane of array irradiance and cell operating temperature for a facade installation in Cabauw, will produce 0.4% higher annual yield than a module with cells optimized for STC. For our module scenario in Alice Springs, the cells with optimized metallization do not significantly increase the module yield since the average irradiance is close to STC. However, with our assumed finger geometry we observe a strong sensitivity for cells with less than approx. 100 fingers with yield losses up to 2.5%. Overall, we find this impact of locally optimized cell metallization is comparatively low, despite the significantly lower average irradiance received by the module.

Two emerging trends in multicrystalline silicon cell texturing are plasma texturing, and metal catalyzed chemical etching. Both processes roughen silicon surfaces in order to increase light absorption. These processes are attractive as they are applicable to diamond-wire sawn wafers. This work investigates the optical properties of these surfaces, and other conventionally textured surfaces like isotropic acidic and random pyramid textures, are investigated for cells in air and after encapsulation for a large range of angles of incidence. We find that the angular optical performance in air varies strongly with cell texture, but when embedded in a module structure these variations are significantly mitigated: the advantages of a high angular absorption of solar cells are not fully transferred to the module level. This is especially notable for plasma etched, black silicon cell structures which suffer comparatively from poorer index matching and light recycling inside a module structure. The losses caused explicitly by the module embedding (described in the cell to module ratio) are in the range of 1–5% for perpendicular incoming light, and increase to 6–15% at an angle of incidence of 70°. Based on these angular performances, we calculate the annual yield of the modules and find that it varies by less than 2% for the cell textures. Nevertheless, the annual optical yield for the black silicon cell structures are the highest, whereas the metal catalyzed chemical etching cell structures show the lowest performance. Further, we find that different annual distributions of the incoming light at the two investigated locations (Melbourne and Alice Springs) only impact the relative performance of the cell textures at high module tilt angles.

New ribbon designs to minimize shading losses are evaluated and compared. With a simulation methodology considering interaction of optics from cells, modules and systems in realistic conditions, the angular responses of modules and annual yield enhancements with these designs are estimated. It is shown that the annual yield cannot be easily estimated from STC measurements and an incident angle multiplier coefficient is proposed for correction.

Cell interconnector designs such as light redirecting films (LRFs) and various geometric ribbon layouts are all aimed at improving the performance of crystalline silicon solar modules. However, due to the usually specular reflecting surface, the angular-dependent module performance, which is typically quantified with an incidence angle modifier, depends on the rotation of the module. We find that typical power measurements under standard test conditions at 1000 W m−2 and 25 °C cell temperature are not suited to identify the best performing design in real-world conditions. In this work, we therefore determine the optimum ribbon geometry based on its simulated energy yield in common installation configurations for solar modules based on industry-standard full and half-cut solar cells. We compare the effects of planar, triangular, LRFs, pentagonal and wire ribbon geometries, as well as fixed optimal inclination, building-integrated (façade), and single-axis tracking installation scenarios of modules in portrait and landscape orientation. Energy yield gains of 1.8% can be achieved for full cell modules with LRFs or pentagonal ribbons in single-axis tracking installation compared to a reference module with five planar ribbons. Critically, we find that changing the module orientation to landscape reduces this energy yield gain by nearly 50% for the modules with LRFs. This directional dependence is significantly reduced with the pentagonal ribbon structure. The installation scenario can have a similarly dramatic impact. The expectation of power gains for certain designs inferred from standard test conditions may not only show significantly lower gains in annual energy yield, but may even lead to yield losses, especially for building-integrated photovoltaic systems. For a typical full-cell module, for example, we have found that LRFs perform 2.5% better than standard planar ribbons under standard test conditions but can result in a 0.2% lower annual yield in a façade installation. Therefore, to fully evaluate the effectiveness of a specific ribbon design, the annual energy yield must consider the angular irradiance distribution and weather conditions at a specific location, the installation scenario, and the module orientation.

High accuracy of yield prediction is of utmost importance for commercial scale photovoltaic systems. One key parameter crucial to the prediction accuracy is the choice and availability of reliable solar radiation data. In this work we investigate the impact of two fundamentally different irradiance data sources on the calculation of the yearly yield and performance ratio for five locations in different climatic regions of Australia. We find an overestimation of the yield calculation of up to 9.3% for satellite-based climate irradiance data compared against one-minute ground-based irradiance data. The yield overestimation shows a general correlation with the number of cloudy days. We propose a linear correction of the yield calculation which allows to improve the prediction accuracy based on more broadly available satellite-based irradiance data.

Solar cells and modules are usually optimized separately under standard test conditions (STC). However, the actual performance under field exposure depends on module design, installation and environmental conditions. This work therefore presents a methodology to predict the annual yield of crystalline silicon solar modules for varying module configurations and environments using a time step simulation approach. We define and quantify 12 effects that affect the annual module performance, starting with solar cells in air under STC to a complete module under realistic conditions. We consider the interaction of optical, thermal and electrical effects in the module and combine it with an angular, spectral and time resolved light source. The model enables understanding the impact of the module environment, such as regions with high wind speeds or high diffuse light, on the annual output of a particular module type. We validate our model with measured power and temperature data for crystalline silicon heterojunction solar cell modules and find an agreement of the annual yield within 0.7%. We quantify the annual gains and losses for this module and identify that major losses are caused by elevated operating temperatures (2.3%) and reflection losses at the front glass (2.6%). Critically, the coupling gains at the cell surface increase significantly by 1.5% when considering angular light irradiance compared to normal incidence. Finally, we apply our Cell-to-Module-Yield method to a 72-cell mono-crystalline PERC module, where the main annual yield losses are caused by reflection at the front glass and ohmic losses in the connector ribbons.

Photovoltaic cell and module manufactures optimise their products according to power measurements performed at a set of standard-test conditions. A key parameter for the financing of a solar project is yield under field or realistic conditions. Field testing modules is time consuming and costly. Hence, we develop a methodology for simulating PV module yield based on the optical, thermal and electrical properties of the components, and the module configuration regarding the cell spacing, interconnection and module layers. With our procedure, we model the performance of standard, half cell and encapsulant free modules in different locations. We present results using our cell to module yield framework for 16 different locations in Australia based on one-minute ground measured solar irradiance and ambient temperature values. We find low-light irradiance losses are directly correlated to the number of cloudy days at a given site. The majority of fielded losses are due to temperature effects, which can be predicted by the average temperature at 3 p.m. We note that wind speed is not accounted for and it will be incorporated in future studies.

Minimizing the optical and electrical losses from cell to module is essential for highly efficient PV modules. We use metal wrap through (MWT) solar cells with passivated rear surface to integrate them into a PV-module and minimize the cell to module losses. For this purpose, we analyze the optical properties of different encapsulation materials with respect to this specific cell type, i.e. the absorption losses in the encapsulants and coupling gains from refractive index. The best performing encapsulant shows a cell-to-module power loss of 0% for a 16-cell MWT module. The fill factor loss is 1.4% absolute. We reach a module efficiency of 17.4% on the aperture area implying a CTM-loss in efficiency of only 0.5% absolute.

PV cell and module manufactures optimise their products according to standard test conditions. The key parameter for financing of a solar farm is yield under field or realistic conditions. Field testing modules is expensive and time consuming. Hence we develop a methodology for simulating PV module yield based on the optical, thermal and electrical properties of the components and their stack ands layout. With our procedure we will model optical, thermal and electrical losses under realistic conditions for standard, half cell and encapsulant free modules in different locations. For now we quantify the losses for a standard module installed in Melbourne on a cloudy day. The largest loss factor is electrical, as the module voltage decreases with low irradiance.

Solar cells made from multicrystalline silicon (mc-Si) wafers play an important role in photovoltaics. Nevertheless, tools for the optical simulation of these devices are scarce. In the present work, the reflectance and charge carrier generation of mc-Si cells and modules are for the first time simulated successfully in the complete spectral range including light trapping and escape light, as the comparison with measured reflectance of the finished cells and mini-modules shows. The “spherical caps” geometry is used to model the front surface reflection of iso-textured silicon solar cells. The characteristic angles of the spherical caps are determined from the reflectance of iso-textured wafers for three different texture strengths. Based on this calibration, the reflectance and charge carrier generation rates of cells encapsulated with EVA and glass are simulated and analysed. Iso-textured cells with full-area aluminium back surface field (Al-BSF) and with passivated emitter and rear (PERC) are quantitatively compared regarding the photo-generated current density jPh. The simulations demonstrate that the direct cell-to-module loss of iso-textured mc-Si cells with Al-BSF (0.7 mA/cm2) is smaller than for PERC cells (1.2 mA/cm2).

The power output and efficiency of a crystalline silicon PV module is a highly relevant quantity to evaluate novel cell, material and design technologies. It is therefore necessary to develop a full understanding of the different factors that impact the cell-to-module ratio. The purpose of this work is to demonstrate a unified methodology for predicting module power and efficiency from material and geometry data for several cell, interconnection and module designs. The methodology consists of analyzing photovoltaic inactive module areas, optical effects such as reflection, absorption and electrical losses. For a selection of efficiency impacting factors we show experimental results to demonstrate their influence on module efficiency and module power. The model has been validated by measurements on an H-patterned multi-crystalline 60-cell module showing a deviation in module power of only 1%, a value well within the flasher measurement uncertainty. For a 60-cell-module with pseudo-square monocrystalline solar cells with 19.9% efficiency we calculate absorption losses for different encapsulants in the range of 1.3 to 5.3 W. The power gain for the investigated backside materials is found to be in the range between 4.5 and 6.7 W.