Advances in Solar Measurement and Modeling at NREL-Manajit Sengupta-NREL

Advances in Solar Measurement and Modeling at NRELDr. Manajit SenguptaDr. Yu XieAron HabteDr Christian Gueymard1 1 th PVPMC Workshop, Wehai, China December 4 5 , 2 0 1 8 NATIONAL RENEWABLE ENERGY LABORATORY 2NREL | 2 2 4 8 weather stations with 2 6 Solar measurement stations [ERDA, NOAA, 1 9 7 9 ] 2 3 9 modeled stations with 5 6 partial measurement stations [DOE, NOAA, 1 9 9 4 ] 1 ,4 5 4 modeled locations[DOE, SUNY-A, NOAA, 2 0 0 7 ] Satellite-based, gridded, 4 km x 4 km, half-hourly [DOE, NOAA, UW, SCS 2 0 1 6 ] http//nsrdb.nrel.gov SOLMET 1 9 7 7 –8 0 NSRDB 1 9 6 1 –1 9 9 0 NSRDB 1 9 9 1 –2 0 0 5 NSRDB 1 9 9 1 –2 0 1 01 ,4 5 4 modeled locations[DOE, CPR, 2 0 1 2 ] NSRDB 1 9 9 8 –2 0 1 7 The National Solar Radiation Database Sengupta, M., Y. Xie, A. Lopez, A. Habte, G. Maclaurin, and J. Shelby 2 0 1 8 , The National Solar Radiation Data Base NSRDB, Renew. Sustain. Energy Rev., 8 9 , 5 1 -6 0 . https//doi.org/1 0 .1 0 1 6 /j.rser.2 0 1 8 .0 3 .0 0 3 NREL | 3 Physical Solar Model PSM Framework Spectral Datasets from the NSRDB NREL | 5 Spectral Data in the Plane-of-ArrayNSRDB Variables Global horizontal irradiance GHI Direct normal irradiance DNI Diffuse horizontal irradiance DHI Clear-sky GHI, DNI, and DHI Cloud type Dew point** Air temperature* Atmospheric pressure Relative humidity** Solar zenith angle Precipitable water*  Wind direction** Wind speed.** Spectral POA 2 0 0 2 wavelengths* From MERRA-2** Recalculated from MERRA-2 NREL | 6 FARMS-NIT for Clear SkyFast All-Sky Model for Solar Applications – Narrowband Irradiance on Tilted Surfaces FARMS-NITSMARTS – Simplified Model of Atmospheric Radiative Transfer of Sunshine.Provides atmospheric properties including atmospheric optical depth, aerosol optical depth, asymmetry parameter and single-scattering albedo NREL | 7 FARMS-NIT for Clear Sky Spectral radiances are computed by solving the radiative transfer equation with the single-scattering approximation for three individual photon paths. The atmospheric radiances are given by radiances related to the three photon paths. POA irradiances are efficiently computed for 2 0 0 2 wavelength bands 0 .2 8 -4 .0 mm from the radiances. Radiances are computed for 4 5 0 sky-view angles that can be integrated for any tilt-geometryAerosolAerosol Aerosol Xie, Y., Sengupta, M., 2 0 1 8 . A Fast All-sky Radiation Model for Solar applications with Narrowband Irradiances on Tilted surfaces FARMS-NIT Part I. The clear-sky model. Sol. Energy, 1 7 4 C, 6 9 1 -7 0 2 .Two-layer model A S SS S A NREL | 8 Evaluation of FARMS-NIT for Clear Sky To validate FARMS-NIT, we use measurements of GHI and cloud fraction at NREL’s SRRL to identify clear-sky conditions shadows. Measurements of precipitable water vapor PWV, aerosol optical depth AOD, and surface albedo are used by the models. Measurements from EKO-WISER spectroradiometer MS-7 1 1 and MS-7 1 2 on a 1 -axis tracker is compared with FARMS-NIT and TMYSpec parameterized model, Myers, 2 0 1 2 . Snow surfaceVegetated SurfaceOct 2 0 , 2 0 1 7 Jan 2 2 , 2 0 1 8 NREL | 9 Evaluation of FARMS-NIT FARMS-NIT has a much better performance than TMYSPEC, especially on the snow day when validated with spectral measurements from the EKO MS-7 1 1 Spectroradiometer. FARMS-NIT slightly overestimates spectral radiation in the UV and visible regions while TMYSPEC underestimates it. FARMS-NIT Mean Bias Error MBE 1 and Absolute Mean Bias Error 4 .Vegetated Surface Snow Surface NREL | 1 0 FARMS-NIT for Cloudy-SkySMARTS provides atmospheric optical depth for layers below and above cloud.Aerosols are not important in cloudy sky situation. NREL | 1 1 FARMS-NIT for cloudy-sky conditions Spectral radiances are computed by solving the radiative transfer equation. Two additional photon paths are considered for Rayleigh scattering under the clouds. Three-layer modelA AS A S S S AAAS S SAA A R SS A NREL | 1 2 FARMS-NIT for cloudy-sky conditions NREL | 1 3 Computing time for FARMS-NIT For computing hourly spectral POA irradiances for a day, the 6 4 -stream DISORT, 1 6 -stream DISORT, FARMS-NIT, and TMYSPEC consume 1 8 0 hours 4 8 minutes, 3 hours 1 8 minutes, 2 1 .9 seconds, and 2 .3 1 seconds. Our current server uses multiple-processors and we can compute and deliver spectral data for 1 year in 2 minutes. Estimating Ultraviolet Radiation from Total Radiation NREL | 1 5 Why UV and How do we Estimate itWhy do we need UV estimates Terrestrial ultraviolet UV radiation is a primary factor contributing to degradation and reliability of materials over time. There is limited availability of UV measurements.How do we estimate UV Measured and/or modeled total solar irradiance TS 2 8 0 –4 0 0 0 nm is relatively abundant. Estimate the clear-sky terrestrial UV irradiance 2 8 0 –4 0 0 nm and 2 8 5 -3 8 5 nm from TS. Develop a model of the UV/TS ratio using simulations obtained with the Simple Model of the Atmospheric Radiative Transfer of Sunshine SMARTS. NREL | 1 6 Goal Worldwide Application The goal is to make the draft ASTM standard representative of all locations around the world. NREL | 1 7 Ruv as a Function of Airmass NREL | 1 8 LOCATION INFORMATION AND ASSOCIATED NUMERICAL COEFFICIENTS OBTAINED BY LEAST-SQUARES FITTING 2 8 0 –4 0 0 nmStation Lat Long Elevationm Numerical Coefficientsm4 m3 m2 m1 m0Birdsville, Australia -2 5 .9 1 3 9 .3 4 6 1 .7 9 E-0 6 -8 .3 9 E-0 5 1 .4 7 E-0 3 -1 .0 1 E-0 2 7 .0 9 E-0 2CEI Qiong Hai, HaiNan province, China 1 9 .2 1 1 0 .5 6 2 2 .8 4 E-0 6 -1 .2 7 E-0 4 1 .9 5 E-0 3 -1 .1 1 E-0 2 7 .0 5 E-0 2CEI Turpan, XinJiang province, China 4 2 .9 8 9 .8 1 0 3 .2 5 E-0 6 -1 .4 5 E-0 4 2 .2 2 E-0 3 -1 .2 2 E-0 2 6 .8 6 E-0 2Case Western Reserve University CWRU, Ohio, USA 4 1 .5 -8 1 .6 2 0 0 2 .5 3 E-0 6 -1 .1 5 E-0 4 1 .8 7 E-0 3 -1 .1 8 E-0 2 7 .0 5 E-0 2Fairbanks, AK, USA 6 4 .8 -1 4 7 .7 1 3 6 1 .0 4 E-0 6 -6 .0 1 E-0 5 1 .2 6 E-0 3 -9 .9 8 E-0 3 7 .7 6 E-0 2KACST Riyadh, Saudi Arabia 2 4 .9 4 6 .4 7 4 0 3 .3 0 E-0 6 -1 .4 6 E-0 4 2 .1 7 E-0 3 -1 .1 6 E-0 2 7 .0 2 E-0 2Miami, Florida, USA 2 5 .6 -8 0 .5 3 0 2 .3 0 E-0 6 -1 .0 9 E-0 4 1 .8 2 E-0 3 -1 .1 5 E-0 2 7 .2 6 E-0 2Nauru -0 .5 1 6 6 .9 7 1 .4 6 E-0 6 -7 .5 2 E-0 5 1 .3 8 E-0 3 -9 .7 6 E-0 3 7 .3 8 E-0 2 NREL-Golden, Colorado, USA 3 9 .7 -1 0 5 .2 1 7 9 0 1 .9 7 E-0 5 -5 .3 9 E-0 4 5 .2 6 E-0 3 -2 .1 8 E-0 2 7 .9 6 E-0 2Petrolina, Brazil -9 .4 -4 0 .5 3 7 0 1 .7 3 E-0 6 -8 .5 3 E-0 5 1 .5 2 E-0 3 -1 .0 4 E-0 2 7 .2 6 E-0 2Phoenix, Arizona, USA 3 3 .9 -1 1 2 .2 3 9 5 1 .9 7 E-0 6 -9 .4 1 E-0 5 1 .6 2 E-0 3 -1 .0 8 E-0 2 7 .0 9 E-0 2Pretoria, South Africa -2 5 .8 2 8 .3 1 4 4 9 2 .9 1 E-0 6 -1 .2 8 E-0 4 2 .0 4 E-0 3 -1 .2 7 E-0 2 7 .0 7 E-0 2Sanary, France 4 3 .1 5 .8 1 1 0 2 .5 0 E-0 6 -1 .1 4 E-0 4 1 .8 6 E-0 3 -1 .1 8 E-0 2 6 .9 7 E-0 2Singapore 1 .3 1 0 3 .8 3 0 3 .1 0 E-0 6 -1 .3 7 E-0 4 2 .0 9 E-0 3 -1 .1 9 E-0 2 7 .1 2 E-0 2Toravere, Estonia 5 8 .3 2 6 .5 7 0 2 .1 6 E-0 6 -9 .9 2 E-0 5 1 .6 7 E-0 3 -1 .1 0 E-0 2 6 .8 4 E-0 2 where AMi is the airmass, and mi are numerical coefficients obtained by least-squares fitting.Fourth-Order Polynomial Functions NREL | 1 9 Variability in UV Estimates at Various Locations 0.00E005.00E-021.00E-011.50E-012.00E-01 2.50E-013.00E-013.50E-014.00E-01 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15R_UV SIte Number UV Estimates for Various Air-masses for all locationsSeries1 Series2 Series3 Series4 Series5 Series6 NREL | 2 0 Modeled vs. measured 1 -min UV global irradiance under all sky conditions at SRRL for low and high surface albedo conditions.The correlation between the modeled and measured UV irradiance is highly significant R 2 0 .9 9 5 , which provides confidence in the model developed here.Modeled vs. measured 1 -min UV global irradiance under clear-sky winter conditions at SRRL. Validation using 1 -minute MeasurementsUV radiometers Eppley Lab TUVR and Kipp Zonen CUV4 NREL | 2 1 Hourly modeled vs. measured UV global irradiance under clear- and cloudy-sky conditions at SRRL for one year August 2 0 1 6 to August 2 0 1 7 .Most of the hourly differences are within 2 W/m2 . There are only a few outliers outside of the range of 4 W/m2 , which could be related to unusual combinations of atmospheric conditions or radiometer maintenance issues.Validation using Hourly Averaged Measurements NREL | 2 2 Development of ASTM StandardASTM Work Item WK5 7 7 1 4 Standard estimation of UV irradiancehttps//ieeexplore.ieee.org/stamp/stamp.jsptparnumber8 5 2 9 2 2 9 https//www.astm.org/DATABASE.CART/WORKITEMS/WK5 7 7 1 4 .htm Low-Cost Multiparameter Sensor for Solar Resource Applications NREL | 2 4 Arable Mark Device Shortwave 400 -700 nm Shortwave 4 0 0 -7 0 0 nmSix-band spectrometer Six-band spectrometer4 -Way NetRadiometerAir TemperatureHumidityPressure Cellular, Wi-Fi and BluetoothInternal Antennas Auxiliary SensingSoil moisture, cameraSolar Power GPS, Tilt, OrientationCellular, Wi-Fi and BluetoothRainfall HailDrop size distributionLongwave radiometerCrop and sky Temperature NREL | 2 5 Arable Mark Reference DataAll-sky comparison at 1 -minute resolutionshows good agreement compared with reference data. Characterization Results NREL | 2 6 A fast spectral POA model was built, validated and implemented to provide on demand spectral radiation from the NSRDB. A model was developed to estimate the GUV irradiance in two different wavebands 2 8 0 –4 0 0 nm and 2 8 5 –3 8 5 nm using the total broadband solar irradiance. The atmospheric airmass was found to be the primary driver of the GUV/TS ratio, at least under “typical” atmospheric conditions. The model does not appear to be significantly affected by cloudiness. The model typically under- or overestimates the measured UV irradiance by only 2 W/m 2 on an hourly basis during the course of one year. We characterized a low cost device for irradiance measurement and showed that it held significant promise for PV applications. Conclusions and Future Work www.nrel.govThis work was authored by Alliance for Sustainable Energy, LLC, the manager and operator of the National Renewable Energy Laboratory for the U.S. Department of Energy DOE under Contract No. DE-AC3 6 -0 8 GO2 8 3 0 8 . Funding provided by U.S. Department of Energy Office of Solar Energy Technology Office. The views expressed in the article do not necessarily represent the views of the DOE or the U.S. Government. The U.S. Government retains and the publisher, by accepting the article for publication, acknowledges that the U.S. Government retains a nonexclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this work, or allow others to do so, for U.S. Government purposes. Thank YouContact Manajit.Senguptanrel.gov NREL | 2 8 Validation at Various Locations NREL | 2 9 Validation NREL | 3 0 Comparison Under Different UV Spectral RangesComparison of results using different definitions of UV spectral rangeStation NREL Model 2 8 0 –4 0 0 nmMJ/m2 * NREL Model 2 9 5 –4 0 0 nmMJ/m2 * Poliskie, 2 0 1 12 9 5 –4 0 0 nmMJ/m2 NREL Model2 8 5 –3 8 5 nmMJ/m2 NREL Model2 9 5 –3 8 5 nmMJ/m2 * White et al., 2 0 1 12 9 5 –3 8 5 nmMJ/m2Case Western Reserve Univ. CWRU, Ohio, USA 2 9 1 0 tilt2 8 5 5 tilt2 6 9 4 1 tilt 2 8 8 0 tilt 2 8 5 5 tilt 2 6 9 4 1 tilt 2 2 7 0 tilt 2 2 1 5 tilt 2 0 8 4 1 tilt 2 2 4 0 tilt 2 2 1 5 tilt 2 0 8 4 1 tilt Miami, Florida, USA 4 2 2 0 tilt 4 1 0 5 tilt 4 0 0 2 6 tilt3 6 9 4 5 tilt 4 1 6 0 tilt4 1 0 5 tilt4 0 0 2 6 tilt3 6 9 4 5 tilt 3 9 0 2 6 tilt 3 3 0 0 tilt3 2 0 5 tilt3 0 4 2 6 tilt2 9 5 4 5 tilt 3 2 5 0 tilt3 2 0 5 tilt3 1 1 2 6 tilt2 8 8 4 5 tilt 3 3 8 5 tilt 3 2 0 4 5 tilt NREL, Golden, Colorado, USA 3 4 1 0 tilt3 4 1 5 tilt3 3 7 4 0 tilt 3 3 9 0 tilt3 4 1 5 tilt3 3 7 4 0 tilt 2 6 6 0 tilt2 6 5 5 tilt2 6 0 4 0 tilt 2 6 4 0 tilt2 6 5 5 tilt2 6 0 4 0 tilt Phoenix, Arizona, USA 4 3 9 0 tilt 4 3 5 5 tilt4 3 2 3 4 tilt 4 3 6 0 tilt 4 3 5 5 tilt 4 3 2 3 4 tilt 4 4 0 3 4 tilt 3 4 3 0 tilt3 3 9 5 tilt3 6 1 3 4 tilt 3 4 0 0 tilt3 3 9 5 tilt3 3 6 3 4 tilt 3 5 9 5 tilt3 6 3 3 4 tilt * Values are obtained using the NREL TMY data set PSM V3 .Note Orientation is south facing