Development of an optical model forsimulating energy yield of a bifacial
PV array
25-26.10.2017
Dimitrij Chudinzow,
Ludger Eltrop
Picture: http://www.desertmodule.cl/
IER Universität Stuttgart 226.10.2017
1. Introduction into bifacial PV
2. Methodology for energy yield modelling
3. Results
Agenda
IER Universität Stuttgart 326.10.2017
… with monofacial PV (!)
PV electricity market prices set new world record in 2016 in Chile
1: https://www.pv-magazine.com
2: https://www.pv-magazine-latam.com
Imagine, how much more efficient could PV become
(even in less sunny regions than Chile)
if using both sides of the module!!!
Introduction ResultsMethodology
IER Universität Stuttgart 626.10.2017
Utility-scale bifacial power plants
• Pel= 1.25 MW
• Inauguration: 2013
• Fixed tilt
1: Photovoltaic Technical Solutions Presentation, 2016
2: isc-konstanz.de
3: pv-magazine.com
• Pel= 2.5 MW
• Inauguration : 2016
• Fixed tilt
La Hormiga PV Power Plant (commercial), Chile2
Sunpreme PV Power Plant (commercial), USA3
• Pel= 12.8 MW
• Inauguration : 2016
• Fixed tilt
Hokuto PV Power Plant (test facility), Japan1
Introduction ResultsMethodology
Bifacial Gain = EnergyrearEnergyfront
IER Universität Stuttgart 726.10.2017
Which bifacial gain can one expect from bifacial PV plants?
Hokuto bifacial PV Power Plant, fixed-tilt, 1.25 MW1
1: World First Large Scale 1.25MW Bifacial PV Power Plant on Snowy Area in Japan, 3rd bifi PV workshop in Miyazaki, Japan, 2016
Introduction ResultsMethodology
IER Universität Stuttgart 826.10.2017
These factors directly influence the
shading constellation and thus the
ground-reflected irradiation from
DNI & DHI (albedo)
SunEdison PV Power Plant, Chile1
Influencing factors on absorbed irradiation
1: pv-magazine.com
1. Location
• Weather conditions
• Ground albedo factor
2. Field layout
• Elevation (installation height)
• Orientation
• Row spacing
• Slope
Introduction ResultsMethodology
IER Universität Stuttgart 926.10.2017
Absorbed irradiation by conventional (monofacial) PV
DNI
DHI
Introduction ResultsMethodology
DNI: Direct Normal Irradiation
DHI: Diffuse Horizontal Irradiation
DNIground, albedo
DHIground, albedo
IER Universität Stuttgart 1026.10.2017
Absorbed irradiation by bifacial PV
DNI
DHI
Ground albedo
from DHI
Shaded area:
calculable size &
position
Ground albedo
from DNI + DHIUnshaded area: unknown
size & position
„How can one properly take ground albedo
irradiation (from DNI & DHI) into account?“
Introduction ResultsMethodology
IER Universität Stuttgart 1126.10.2017
Definition of „view fields“
Dashed lines
seperate front
and rear side
Ground albedo:
DNI + DHI
Length
Width
Introduction ResultsMethodology
IER Universität Stuttgart 1226.10.2017
Definition of „view fields“
• Length of front and rear view fields depends on row
spacing, elevation and slope
• FVF of first and RVF of last row are treated differently
• Width of view fields is defined exogeneously:
• Higher widthmore albedo energy (how much
more?), but also higher land purchase costs
For a fixed-tilt configuration, the view fields are
time-invariant
front rear
Length of front view field (FVF) Length of rear view field (RVF)
Introduction ResultsMethodology
Sideview
IER Universität Stuttgart 1326.10.2017
View fields implemented in Matlab
Here no self-shading occurs (7h after sunrise) Here self-shading occurs (12h after sunrise)
Introduction ResultsMethodology
IER Universität Stuttgart 1426.10.2017
View fields implemented in Matlab
Here no self-shading occurs (7h after sunrise) Here self-shading occurs (12h after sunrise)
Introduction ResultsMethodology
IER Universität Stuttgart 1526.10.2017
Recap of developed methodology
Introduction ResultsMethodology
„How can one properly take ground albedo
irradiation (from DNI & DHI) into account?“
Definition of view fields
+
using theory of view factors to compute
share of ground albedo irradiation that hits
the module‘s surfaces
IER Universität Stuttgart 1626.10.2017
Simulation set-up
Introduction ResultsMethodology
• Location: Atacama Desert, Chile
• Weather data in hourly resolution
• Computation resolution: 20 min
• PV Array: 80 modules in 4 rows (≈21 kWel)
• Ground reflectivity: 20%
IER Universität Stuttgart 1726.10.2017
Monthly absorbed irradiation
Introduction ResultsMethodology
Slope=25°, elevation= 3.5m, row spacing=4m
IER Universität Stuttgart 1826.10.2017
Annual absorbed energy
Introduction ResultsMethodology
Slope=25°, elevation=2m, variation of row spacing
IER Universität Stuttgart 1926.10.2017
Annual bifacial gain of absorbed irradiation (AI)
Introduction ResultsMethodology
Row spacing=5m, variation of slope and elevation
Bifacial GainAI =Absorbed IrradiationrearAbsorbed Irradiationfront
IER Universität Stuttgart 2026.10.2017
Conclusions
Introduction ResultsMethodology
• The definition and implementation of „view fields“ for both sides of a module is
suitable to take into account different irradiance contributions
• Bifacial gain highly depends on array desing
(slope, elevation, row spacing, assumed width of view fields)
• Simulations show that a bifacial gain of over 40% of absorbed energy is possible
• Next step is to implement a submodel to calculate produced electric energy
Muchas Gracias!
Telefon +49 (0) 711 685-87870
Fax +49 (0) 711 685-87873
Universität Stuttgart
Heßbrühlstraße 49a
70565 Stuttgart
Dimitrij Chudinzow, M. Sc.
IER Institut für Energiewirtschaft
und Rationelle Energiewendung
IER Institut für Energiewirtschaftund Rationelle Energieanwendung
Annex
IER Universität Stuttgart 2326.10.2017
𝜙12 =diffuse energy leaving A1directly toward and intercepted by A2
total diffuse enrgy leaving A1
1
𝜙12 =1
𝜋𝐴1𝐴1
𝐴2cos 𝛽1 ∙cos(𝛽2)
𝑠²𝑑𝐴1𝑑𝐴2
1
Theory of „view factors“
1: Heat Transfer Handbook, 2003
2: VDI Wärmeatlas, 2013
Theory of „view factors“ 2
Assumption: All surfaces radiate diffusely
View factor is a solely geometric quantity
Sum of all view factors for one surface is always 1
(conservation of energy)
Introduction ResultsMethodology
IER Universität Stuttgart 2426.10.2017
• The time-variant view factors from „ground shadowmodule row“ and „view fieldmodule row“
are calculated using an algorithm in Matlab, developed by Nicolas Lauzier1
• Time-invariant view factors „module row sky“ (Perez model) are calculated using methodologies
from:
• For inner rows: View factors of photovoltaic collector systems, Maor, T.; Appelbaum, J., 2012
• For outer rows: FC→sky =1+cos(𝛽)
2, 𝛽 = 𝑚𝑜𝑑𝑢𝑙𝑒′𝑠 𝑠𝑙𝑜𝑝𝑒
(Both approaches for time-invariant view factor calculation assume infinitely long module rows)
Calculation of „view factors“
1: https://de.mathworks.com/matlabcentral/fileexchange/5664-view-factors
Introduction ResultsMethodology
IER Universität Stuttgart 2526.10.2017
Daily absorbed energy
Introduction ResultsMethodology
San Pedro de Atacama, slope=25°, elevation=3.5m, row spacing=4m
IER Universität Stuttgart 2626.10.2017
Annual absorbed energy
Introduction ResultsMethodology
Slope=25°, elevation=3.5m, row spacing=4m
IER Universität Stuttgart 2726.10.2017
Low Solar Bids
https://c1cleantechnicacom-wpengine.netdna-ssl.com/files/2016/08/low-solar-bids.png
IER Universität Stuttgart 2826.10.2017
Annual bifacial gain of absorbed energy
Introduction ResultsMethodology
Row spacing=4m, variation of slope and elevation
Bifacial GainAI =Absorbed IrradiationrearAbsorbed Irradiationfront