Spectral derates phenomena of atmospheric components on multi-junction CPV technologies

Abstract

Abstract. The solar spectrum varies with atmospheric conditions and composition, and can have significant impacts on the output power performance of each junction in a concentrating solar photovoltaic (CPV) system, with direct implications on the junction that is current-limiting. The effect of changing solar spectrum on CPV module power production has previously been characterized by various spectral performance parameters such as air mass (AM) for both single and multi-junction module technologies. However, examinations of outdoor test results have shown substantial uncertainty contributions by many of these parameters, including air mass, for the determination of projected power and energy production. Using spectral data obtained from outdoor spectrometers, with a spectral range of 336nm-1715nm, this investigation examines precipitable water (PW), aerosol and dust variability effects on incident spectral irradiance. This work then assesses air mass and other spectral performance parameters, including a new atmospheric component spectral factor (ACSF), to investigate iso-cell, stacked multijunction and single-junction c-Si module performance data directly with measured spectrum. This will then be used with MODTRAN5 ® to determine if spectral composition can account for daily and seasonal variability of the short-circuit current density J sc and the maximum output power P mp values. For precipitable water, current results show good correspondence between the modeled atmospheric component spectral factor and measured data with an average rms error of 0.013, for all three iso-cells tested during clear days over a one week time period. Results also suggest average variations in ACSF factors with respect to increasing precipitable water of 8.2%/cmH 2 O, 1.3%/cmH 2 O, 0.2%/cmH 2 O and 1.8%/cmH 2 O for GaInP, GaAs, Ge and c-Si cells, respectively at solar noon and an AM value of 1.0. For ozone, the GaInP cell had the greatest sensitivity to increasing ozone levels with an ACSF variation of 0.07%/cmO 3 . For the desert dust wind study, consistent ACSF behavior between all iso-cells and c-Si was found, with only significant reductions beyond 40mph.

Low-Cost Spectral Sensor Development Description.

“…Absorption bands for ozone and other atmospheric gasses were not shown to have large impacts on the solar spectrum throughout the course of the day. Further corroboration of this result can be found in [1], where the investigators only found a 0.07%/cmO 3 change in module energy generation. In our low-cost sensor development we limited to the number of spectral bands to five, since the addition of spectral bands would increase the cost of the device.…”

Determination of spectral variations by means of component cells useful for CPV rating and design

“…Several authors have analyzed the effect of spectral variation on the annual current produced by a MJ solar cell in a particular location using some representative spectra or radiative models as MODTRAN , SPCTRAL2 or SMARTS . However all these models have the limitation that they use as inputs the atmospheric parameters whose measurement is complex and commonly require expensive instrumentation.…”

“…In particular, it would be interesting to know how the spectrum evolves throughout the year, what the representative spectrum for a particular location is and how a CPV system could be optimized to maximize the energy harvested in a year. Several approaches have been proposed to account for the spectral influence on the power generation such as the direct measurement of the spectrally resolved direct normal irradiance (DNI) by means of spectroradiometers , the atmosphere characterization based on sun photometers which provides significant parameters like aerosols content and precipitable water (PW) , the atmosphere simulation by radiative transfer models such as MODTRAN , SMARTS or SPCTRAL2 , the analysis of fill factor (FF) variations as a mean to determine the influence of spectrum under operating conditions or the use of component (or isotype) cells and the associated parameters (Z parameter , or spectral matching ratio (SMR) ). Additional parameters proposed to evaluate spectral distribution of the DNI are average energy per photon , useful fraction , spectral factor and absolute air mass (AM) corrected with altitude .…”

Contributions to characterization, modeling and design of concentrator photovoltaic systems

“…Several authors have analyzed the effect of spectral variation on the annual current produced by a MJ solar cell in a particular location using some representative spectra McMahon et al 2002) or radiative models as MODTRAN (Armijo et al 2014), SPCTRAL2 (Araki and Yamaguchi 2003;Faine et al 1991) or SMARTS (Jaus and Gueymard 2012;Lumb et al 2010). …”

“…We propose an approach that involves the use of atmospheric radiative transfer models (Chan, Brindley, and Ekins-Daukes 2014), such as SMARTS , SPECTRAL2 (Faine et al 1991), and MODTRAN (Armijo et al 2014), which synthesize spectrally resolved DNI and GNI (Global Normal Irradiance) upon a particular atmospheric and environmental condition. The spectrally-resolved DNI is obtained by starting with the extraterrestrial irradiance, corrected by actual sun-earth distance, and considering atmospheric factors with a known impact on transmittance, namely, Rayleigh scattering, absorption by ozone, nitrogen dioxide, uniformly mixed gases and water vapor, and aerosol extinction.…”