Search

This study explores the thermal behavior of Building Integrated Photovoltaics (BIPV) façades, focusing on the impact of different mounting configurations and colored interlayers. Two experimental setups are analyzed to investigate the operating temperatures of PV modules. The first setup consists of single-cell mini-modules including both colored and uncolored modules operating under ventilated and insulated conditions. The Sandia model is utilized to correlate cell temperatures and module backside temperatures, for which new empirical I-T coefficients are determined for the different configurations and colored interlayers. The second setup consists of a BIPV container with a south-facing 16-module BIPV curtain wall façade. The modules are mounted with different air gap sizes to examine the effects of rear ventilation on module temperature. The results highlight the significant impact of mounting design and display the complicated nature of temperature distribution of BIPV systems. This research aims to contribute to a broader understanding of the influence that BIPV system integration in architecture has on module operating temperatures.

While colored photovoltaics are gaining popularity in the market for building-integrated photovoltaics (BIPV), several specific properties are not accounted for in standard PV performance models. This work shows how relying on the coloration efficiency alone can lead to significant errors regarding module temperatures. By comparing measured temperature data from a test installation featuring BIPV façade elements in multiple colors, little correlation is found between total optical losses (reflection and absorption losses) and module temperature. Instead, better correlation is found with total reflectance. This is attributed to the light absorbed in the pigment-based colored layers contributing to module heating, whereas reflected light does not. This is especially relevant for colors with high lightness, such as gray or beige, for which reflection losses are dominating absorption losses. When modelling colored BIPV products, it is therefore recommended to only consider reflection losses for the irradiance contributing to module heating, while continuing to also include absorption losses for the effective irradiance used in electrical performance modelling.

The aesthetic challenge in deploying building-integrated photovoltaics (BIPV) is addressed by exploring coloration solutions, particularly absorptive materials. These, aside from the direct coloration losses, also introduce additional absorption and reflection losses, which increase with the incidence angle and vary significantly depending on the color. In this study, BIPV modules with colored interlayer foils are investigated, revealing significant differences in incidence angle modifier (IAM) behavior between the five colors studied. The Martin and Ruiz model captures the overall trend but faces difficulties in fitting the IAM of colored samples, especially when combined with satinated or structured glass. In contrast, the polynomial Sandia IAM model achieves a solid fit. Furthermore, this study compares the lab-determined angular optical losses with outdoor measurement data. This involves modeling effective irradiance and power output, accounting for color-dependent module efficiency, temperature effects, and IAM losses. The study also indicates that apart from IAM losses, spectral mismatch variation has a significant effect on the performance of colored BIPV.

Satinated glass is an increasingly important component of PV modules, especially in the building integrated (BIPV) sector, where it can contribute to more uniform appearance as well as significantly reduce the risk of glare. Different satination methods, such as wet chemical etching or mechanical abrasion, are available, which can impact the PV-relevant optical parameters in different ways. In order to characterize the associated performance losses, in this study eight different satinated glass types are compared in terms of spectral transmittance, surface morphology, relative transmission loss and angular dependent reflection losses. Measurements show only insignificant transmission losses of < 1 % for acid-etched glass compared to reference PV glass at normal incidence, with significant gains at high incidence angles > 60°. No significant differences between different acid-etched samples can be observed, all featuring homogeneous appearance. Other satination technologies, such as sandblasting or laser-etching show significantly higher transmission and reflection losses at low incidence angles, making them less viable for use in PV applications.

Potential glare through reflected sunlight can be a significant hindrance factor for PV installations, especially for building-integrated photovoltaics (BIPV) with atypical orientations and tilt angles present. Structured glass surfaces or antireflective (AR) coatings are often used as a solution, however there currently is no universally recognized method or metric for estimating their impact on glare. This work presents an evaluation of the glare potential of different surface glasses for use in PV modules based on their bi-directional reflectance distribution functions (BRDF). BRDF threshold values for retinal burn damage, flash blindness and discomfort glare are calculated based on retinal irradiance thresholds from literature. Subsequently, gonioreflectometric measurements on PV mini-modules are used to characterize the reflectance profiles of eight different glass surfaces. Results for all measured structured glasses, both satinated and macro-textured, show significant reductions in BRDF compared to smooth glass, largely eliminating the potential for flash blindness as well as discomfort glare at low incidence angles. At high incidence angles, increased potential for discomfort glare as well as forward scattering along the glass surface can be observed. AR coatings, however, are shown to be insufficient to eliminate glare causing flash blindness.

Measurements of the incidence angle modifier (IAM) of colored PV modules based on inkjet-printed glass have shown differences between colors, which are uncorrelated with the total transmittance and reflectance of the colored glass. In this paper, the complex refractive index of ceramic inks is estimated based on measurements of reflectance and transmittance spectra. These estimates are used in ray-tracing simulations to determine the IAM of coupon-sized PV modules with printed ceramic inks. Comparisons of the simulation results to measured IAMs show a high correlation between the refractive index of colorants and the IAM. Furthermore, estimations of the impact of IAM on performance based on TMY data at different locations are performed. The results indicate that changes in IAM due to sample coloration can lead to annual irradiance losses of up to 5% compared to uncolored glass. These losses are caused by additional reflections at the glass-ink interface according to the Fresnel equations, dependent on the refractive index of the ceramic inks.

In this work we investigate the characteristics of solar cells cracks in photovoltaic (PV) modules for understanding the extent to which the solar cell electrical parameters change due to cell crack degradation. The experimental investigation is performed on two custom nine-cell mini-modules of mono- and multi-crystalline silicon, respectively, where each solar cell in the module has a junction box, allowing individual and module level characterization. Results show that power loss caused by cell cracks is driven primarily by a reduction of the cell’s maximum power point current, in particular B type cracks. C cracks also affect the short circuit current of the cells, whereas cells with combined B and C cracks show the most reduction of the short circuit current. Equivalent circuit diode model curve fitting and analysis of the light or dark I-V curves proved of limited used in analyzing degradation of cracked cells, as the model assumptions break down. However, a comparative analysis of dark and light I-V curves with a Suns-Voc curve was better suited for understanding the evolution of diode parameters on cracked cells for increasing levels of degradation.

This contribution reports on a yearlong spectral albedo measurement campaign performed in Roskilde, Denmark. Four albedo scenarios are monitored using three sensor types. The ground surfaces include green grass, dry grass, gravel, and snow – all of which have been monitored with albedometers based on spectroradiometers, silicon-pyranometers, and thermopile pyranometers. Implications of using the various albedo data sources/assumptions in bifacial PV modeling are assessed with the spectrally weighted bifacial energy gain (BEG). We find that BEG differs by as much as 3% using the three albedo sensors and can deviate by 7% from the ground truth when an incorrect static spectral albedo assumption is used. Finally, the spectral mismatch factor (SMM) is calculated to summarize rear plane of array (POA) spectral shifts. Our measurements show midday backside POA spectral shifts as high as 25% for Silicon bifacial PV devices mounted on single axis trackers above grass.

This paper analyzes 15-months of spectral albedo measurements collected at the Technical University of Denmark (55.6°N, 12.1°E). High-resolution spectroradiometers are used to monitor four albedo scenarios, which include green vegetation, dry vegetation, gravel, and snow. Spectral mismatch and spectral impact are calculated for the front and backside of three different bifacial cell concepts mounted on horizontal single axis trackers and fixed-tilt substructures. The spectral nature of albedo is shown to have significant influence on bifacial photovoltaic performance wherein backside spectral impact as high as 1.20 is observed for fixed-tilt systems above green vegetation and as low as 0.98 for systems above snow. The results reveal that spectral impact is always lower on tracked than fixed-tilt systems because a greater fraction of sky diffuse light reaches the backside of tracked systems. Given the variety of albedos tested here, we find that the normalized difference vegetation index is a good predictor of backside spectral effects. When the high-resolution measurements are truncated to 4 to 8 carefully selected wavelengths, we find that this limited measurement resolution sufficiently captures the seasonal spectral albedo fluctuations that influence bifacial photovoltaic energy production. Finally, to alleviate the dearth of spectral datasets presently available to the PV community, the spectral irradiance and albedo measurements are made freely available in open access format.

In this paper, we present a study of daylight EL acquisition and the results of a sunlight variation study in a scenario necessary to assure the increase of EL image quality with denoising by averaging for the robustness of the drone system when bright and intermittently cloudy days occur. It was verified that the indicator of image quality based on the signal-to-noise ratio of EL images has a linear behavior with the amount of averaged images when there is no sun variation. When there are sun irradiance variation, it is observed that the quality decrease even with the increased number of images being averaged, turning to increase again only with further additional images.

Electroluminescence (EL) imaging inspections of PV power plants can bring a huge improvement in accuracy. The use of InGaAs camera will also make such inspections fast, but the restriction to acquire the images during dusk or evening is a limitation. Performing lock-in EL is a way to go for daylight EL. This paper proposes an extension of the SNR50 quality measure to estimate the quality of a stack of N images and evaluates the impact of some factors over the measured and visual quality of images acquired with InGaAs sensors. The factors analyzed are the characteristics of the noise in the acquired images, the influence of the sun variations and the averaging over multiple acquired images.

The experimentation with a movable outdoor electroluminescence (EL) testbed is performed in this work. For EL inspections of PV power plants, the fastest scenario will include the use of unmanned aerial vehicle (UAV) performing image acquisition in continuous motion. With this motivation, we investigate the EL image quality of an acquisition in motion and the extent of image processing required to correct scene displacement. The results show processed EL images with a high level of information even when acquired at 1 m/s camera speed and at frame rate of 120 fps.

With the significant growth in the number of photovoltaic (PV) installations and their size, regular PV system inspection has become a challenge. Aerial drone imaging, based on visual, thermographic, and more recently luminescence, can be viable solutions for PV inspection. However, to achieve effective detection and quantification of failure based on images acquired form Unmanned Aerial Vehicle, there is need for image quality enhancement and correction of distortions, inherent to the drone measurement process. In this work we propose methods to automatically correct the perspective distortion in electroluminescent (EL) images of PV panels. We identified two main cases of perspective distortion: When the imaging plane is parallel to the panel plane or not, and propose methods to correct both. For both cases, the proposed method yields good results, as assessed by visual evaluation.

One year of spectrally resolved direct normal irradiance (DNI), diffuse horizontal irradiance (DfHI) and global horizontal irradiance (GHI) measured in Roskilde, Denmark are analyzed in terms of the average photon energy (APE). We show that the APE of the GHI component is characterized by spectral distributions with low standard deviations, which is consistent with the findings of previous authors. In contrast, the APE of the DfHI and DNI components show spectral distributions with higher standard deviations, which suggests that a given APE value for DNI or DfHI can represent a wider array of spectral distributions. Finally, it is shown that the DfHI APE is characterized by high energy spectral distributions, which has particular significance in the case of building integrated photovoltaic (BIPV) applications in Nordic countries.

The experimentation with a movable outdoor electroluminescence (EL) testbed is performed in this work. For EL inspections of PV power plants, the fastest scenario will include the use of unmanned aerial vehicle (UAV) performing image acquisition in continuous motion. With this motivation, we investigate the EL image quality of an acquisition in motion and the extent of image processing required to correct scene displacement. The results show processed EL images with a high level of information even when acquired at 1 m/s camera speed and at frame rate of 120 fps.

This works present a method for synchronization of a solar device to real solar time for event triggering and energy management. Specifications require the devices to be off-grid and self-adjusting. Measurements of daylight duration were performed using one test device over the course of several months in Denmark. Analysis of viability over the accuracy of solar time determination and potential for self-adjusting energy management was carried out. Measurement results allowed for the development of an algorithm with good potential for performing the required tasks. Influence of variations in weather and different shadowing conditions were tested in two prototype devices. Results show a good potential for daily synchronization between devices and estimation of solar time within less than 5 minutes error.

The results of photoluminescence (PL) imaging performed with a laser line scan in a prototype setup are presented in this work. Here we show the PL images with the photovoltaic cell positioned at different distances from the laser line source, therefore with the beam at different intensities and line characteristics. This study is motivated by the time demanding process of acquiring luminescence images for accurate PV inspections in the field, usually made with electroluminescence imaging. The PL imaging induced by a laser line has the potential to be compact enough to be portable and carried out by drones for field inspections.

On-site inspection of PV systems has been historically performed though visual inspection, infrared (IR) thermography, and electrical measurements. Recent advances and cost reductions in unmanned aerial vehicle (UAV)technology has led to adoption of UAVs equipped with thermal cameras for inspection of PV plants,which survey power plants in a fraction of the time and cost than walk through IR imaging. IR imaging however, is limited only to detection of certain fault types that result in elevated temperatures. Techniques such as electro-(EL) and photo-(PL) luminescence imaging offer a higher level of image detail and qualitative insight compared to IR thermography. Furthermore, detection and identification of incipient or severe faults in PV panels is more straightforward.This project proposes for the first time a fast and accurate automatic drone-based inspection method for large PV plants that combines IR, EL, PL imaging, and visual images (VI), called DronEL. The overarching goal is to correlate these images with known PV failures such as hotspots, cell cracks, and potential induced degradation. The DronEL project is carried out by a number of academic and commercial partners including Denmark’s Technical University (DTU), Aalborg University (AAU), Sky-watch, SiCon and Kenergy.

Reflectors to bifacial PV-cells are simulated and prototyped in this work. The aim is to optimize the reflector to specific latitudes, and particularly northern latitudes. Specifically, by using minimum semiconductor area the reflector must be able to deliver the electrical power required at the condition of minimum solar travel above the horizon, worst weather condition etc. We will test a bifacial PV-module with a retroreflector, and compare the output with simulations combined with local solar data.