Lawrence Dunn

Light soaking effects on photovoltaic modules: Overview and literature review

Device performance under extended duration illumination is an essential characterization step for any PV technology, because light exposure can produce a variety of effects which influence the determination of both initial and long-term stabilized device performance. We present an overview of PV light soaking behavior based on a literature review of light soaking effects on commercial PV module technologies, including a-Si/μc-Si, CdTe, CIS/CIGS, and c-Si. We address the physical mechanisms of light-induced changes in each PV technology, short-and long-term light-induced effects, and current literature knowledge on PV module preconditioning for accurate power output determination. We highlight common themes and identify desirable attributes of equipment for PV module light soaking characterization and test.

Comparison of pyranometers vs. PV reference cells for evaluation of PV array performance

As the photovoltaics (PV) industry has grown, the need for accurately monitoring the solar resource of PV power plants has increased. Historically, the PV industry has relied on thermopile pyranometers for irradiance measurements, and a large body of historical irradiance data taken with pyranometers exists. However, interest in PV reference devices is increasing. In this paper, we discuss why PV reference devices are better suited for PV applications, and estimate the typical uncertainties in irradiance measurements made with both pyranometers and PV reference devices. We assert that the quantity of interest in monitoring a PV power plant is the equivalent irradiance under the IEC 60904-3 reference solar spectrum that would produce the same electrical response in the PV array as the incident solar radiation. For PV-plant monitoring applications, we find the uncertainties in irradiance measurements of this type to be on the order of ±5% for thermopile pyranometers and ±2.4% for PV reference devices.

Comparing PV power plant soiling measurements extracted from PV module irradiance and power measurements

The accumulation of dust and other environmental contaminants on PV modules, also known as PV module soiling, is a significant source of lost potential power generation for PV installations. Designers and operators of utility-scale solar power plants are increasingly seeking methods to quantify soiling-related losses, in order to improve performance modeling and verification or to optimize washing schedules. Recently, soiling measurement equipment has been introduced based on the measurement of two co-planar PV modules, one of which is regularly cleaned, and the other of which naturally accumulates environmental contaminants. These measurements are used to determine a soiling ratio (SR), which may be applied as a derate factor in analysis of the PV system performance. In this work, we examine the difference between a soiling ratio metric calculated from measured temperature-corrected short-circuit current values (SRIsc), which represents the fraction of irradiance reaching the soiled modules, versus a soiling ratio calculated from measured temperature-corrected PV module maximum power values (SRPmax), which represents the fraction of power produced by the soiled modules compared to clean modules. We examine both techniques for CdTe and c-Si module technologies. This study is motivated by the fact that variations in module efficiency versus irradiance, as well as any non-uniformity of soiling, may introduce differences between the power losses estimated from short-circuit current values versus actual soiling-induced power losses. For CdTe, the SRIsc method is found to be a good proxy for the SRPmax method for nonuniform soiling levels up to 11%.

PV module soiling measurement uncertainty analysis

The accumulation of dust and other environmental contaminants on PV modules, also known as PV module soiling, is a significant source of lost potential power generation for PV installations. Designers and operators of utility-scale solar power plants are increasingly seeking methods to quantify soiling-related losses, in order to improve performance modeling and verification or to optimize washing schedules. Recently, soiling measurement equipment has been introduced based on the measurement of two co-planar PV modules, one of which is regularly cleaned, and the other of which naturally accumulates environmental contaminants. These measurements are used to determine a soiling ratio (SR), which may be applied as a derate factor in analysis of the PV system performance. In this work, we have performed an analytical calculation of the uncertainty in soiling ratio measurements. We find that soiling ratio measurements can be performed with uncertainties on the order of ~±1% or better on an absolute basis, depending on calibration conditions, operating temperatures, angular alignments, and other factors we discuss.

Light soaking measurements of commercially available CIGS PV modules

CIGS devices are known to exhibit metastabilities and performance changes with continuous light exposure, or light soaking (see Ref. [1], and references therein). Such metastabilities have an impact on the measurement of CIGS PV module performance, since efficiency and other parameters change on a time scale of hours. For this work we used an indoor continuous solar simulator to expose three commercially available CIGS modules from three different manufacturers to a simulated diurnal light exposure cycle for 16 days. We observed an initial increase in efficiency on the order of ~3% to ~5% at the start of each illumination cycle in all three modules. We also observed a a deviation of approximately -0.17%/°C between the measured value and data sheet value of one modules Pmax/efficiency temperature coefficient, indicating that this module have ~5% lower power output at normal operating temperatures that indicated by the datasheet. In a follow-on experiment, the time required for modules to relax in the dark to their low-efficiency states was investigated by varying the length of time spent in the dark. One module, with an overall power conversion efficiency (PCE) of ~9.3%, required between 2 and 3 hours in the dark to relax to its low efficiency state, while the other two modules, with PCEs of ~10.3%, require between 9 and 16 hours to relax to their low-efficiency states.

Performance analysis of photovoltaic installations in a Solar America City

We present our findings from a recent analysis of monitoring data collected on over 480 residential and commercial PV installations in Austin, Texas between 2005 and 2008. These systems were installed under a city rebate program administered by the city-owned electric utility, Austin Energy, and recognized under the Solar America initiative. The majority of these systems were residential installations with rated power under 4kW. In conjunction with the utility, we have undertaken an analysis of the accumulated data in order to quantify long-term performance. The primary analysis goals were to statistically compare the output of existing installations to predictions based on the PVWatts calculator developed by the National Renewable Energy Laboratory (NREL) [1], and to identify trends linking underperforming or overperforming installations. In addition, we wished to establish a simple methodology for the utility to use in assessing the long-term performance of residential and commercial PV installations.

Pyranometers and Reference Cells: Part 2: What Makes the Most Sense for PV Power Plants?; Preprint

As described in Part 1 of this two-part series, thermopile pyranometers and photovoltaic (PV) reference cells can both be used to measure irradiance; however, there are subtle differences between the data that are obtained. This two-part article explores some implications of uncertainty and subtleties of accurately measuring PV efficiency in the field. Part 2 of the series shows how reference cells can be used to more confidently predict PV performance, but how this could best be accomplished if historic irradiance data could be available in PV-technology-specific formats.