Simulation of Fluorescent Concentrators - SCI Utah
M. Bendig, J. Hanika, H. Dammertz, J. C. Goldschmidt, M. Peters, and M. Weber | August 10, 2008
Simulation of Fluorescent Concentrators
Simulation of Fluorescent Concentrators | Fluorescent Concentrators
Page 2
| August 10, 2008
Solar Energy Research I
cooperation with the Fraunhofer Institute for Solar Energy Systems in Freiburg
I
raise efficiency – by concentration of light
I
problem: diffuse light
I
idea (around 1970): trap light inside medium
I
low efficiency → low research interest
I
interest increased again in the last years due to improved dyes, solar cells and concepts
Simulation of Fluorescent Concentrators | Fluorescent Concentrators
Page 3
| August 10, 2008
What is a Fluorescent Concentrator?
I
PMMA (acrylic glass) and fluorescent dye
I
concentrates direct and diffuse light on a solar cell principle:
I
I I I
absorption in dye re-emmitance according to Stokes shift (longer wavelength) total internal reflection (trap light inside medium)
Simulation of Fluorescent Concentrators | Fluorescent Concentrators
Page 4
| August 10, 2008
Absorption and Photoluminescence Spectra Absorption and Photoluminescence Spectra absorption photoluminescence
0.018 0.016
[dimensionless]
0.014 0.012 0.01 0.008 0.006 0.004 0.002 0 300
I
400
500 600 wavelength [nm]
re-absorption only possible in overlap
700
800
Simulation of Fluorescent Concentrators | Fluorescent Concentrators
Page 5
| August 10, 2008
Energy Loss Mechanisms
I
loss mechanisms have to be analysed for improvement
I
experimental testing of new concepts and analysis of physical processes difficult and expensive
I
analytical calculations and simulation complex and time-consuming
Simulation of Fluorescent Concentrators | The Simulation
Page 6
| August 10, 2008
Simulation Model ⇒ Monte Carlo (MC) method I
ray tracing of single photons (simplest simulation model, least error-prone)
I
uni-directional particle transport to avoid problems arising when connecting paths with different wavelengths
I
allows to investigate effects isolated from global path
Simulation of Fluorescent Concentrators | The Simulation
Page 7
| August 10, 2008
A Photon’s Path trough the Concentrator
I
sample spectrum AM 1.5 (MC inversion method)
I
sample start point
I
fix direction
Simulation of Fluorescent Concentrators | The Simulation
Page 8
| August 10, 2008
A Photon’s Path trough the Concentrator
I
calculate intersection with boundary
I
calculate reflection coefficient (Fresnel, Cauchy )
I
calculate new ray direction (Snell)
Simulation of Fluorescent Concentrators | The Simulation
Page 9
| August 10, 2008
A Photon’s Path trough the Concentrator
I
sample pathlength (MC inversion method, Lambert-Beer )
Page 10
Simulation of Fluorescent Concentrators | The Simulation
| August 10, 2008
A Photon’s Path trough the Concentrator
I
estimate absorption event
Page 11
Simulation of Fluorescent Concentrators | The Simulation
| August 10, 2008
A Photon’s Path trough the Concentrator
I
sample PL-spectrum (MC inversion method)
I
sample direction (MC inversion method)
I
sample pathlength (MC inversion method,) Lambert-Beer
Page 12
Simulation of Fluorescent Concentrators | The Simulation
| August 10, 2008
A Photon’s Path trough the Concentrator
I
calculate intersection with boundary
I
calculate reflection coefficient (Fresnel, Cauchy )
I
calculate new ray direction (reflection angle equals angle of incidence)
Page 13
Simulation of Fluorescent Concentrators | The Simulation
| August 10, 2008
A Photon’s Path trough the Concentrator
I
sample pathlength (MC inversion method, Lambert-Beer )
I
estimate absorption event
Page 14
Simulation of Fluorescent Concentrators | The Simulation
| August 10, 2008
A Photon’s Path trough the Concentrator
I
sample PL-spectrum (MC inversion method)
I
sample direction (MC inversion method )
I
sample pathlength (MC inversion method, Lambert-Beer )
I
calculate intersection with boundary
Page 15
Simulation of Fluorescent Concentrators | The Simulation
| August 10, 2008
A Photon’s Path trough the Concentrator
I
estimate absorption event → terminate ray
Page 16
Simulation of Fluorescent Concentrators | The Simulation
| August 10, 2008
Experiments and their Simulation I
physical measurement of input parameters: absorption spectrum, photoluminescence (PL-) spectrum, refractive indices and geometry
I
several experiments for analysis of fluorescent concentrators first step:
I
I I I
I
reproduce experiments “exactly” compare data to verify implementation explain small differences
second step: I I I
simulate experiments without limitations of “real” setup parameter variation evaluate results
Simulation of Fluorescent Concentrators | The Simulation
Page 17
| August 10, 2008
Reflection and Transmission Experiment
experiments with integrating sphere Transmission Spectrum
Reflection Spectrum 100 rays detected in reflection direction [%]
100
transmitted rays [%]
80
60
40
20
0 300
experimental data simulation data 400
500 600 Wavelength [nm]
700
800
experimental data simulation data
80
60
40
20
0 300
400
500 600 wavelength [nm]
700
800
Page 18
Simulation of Fluorescent Concentrators | The Simulation
| August 10, 2008
Absorption Absorption ≈ 1 − Reflection − Transmission calculated absorption as input → verification of simulated absorption Absorption Spectrum 100 rays with at least one absorption [%]
I I
experimental data simulation data
80
60
40
20
0 300
400
500 600 wavelength [nm]
700
800
Page 19
Simulation of Fluorescent Concentrators | The Simulation
| August 10, 2008
Angular Experiment Blind
PMMA half cylinder Light Photonic structure (optional) Detector Mirror (optional) Optical coupling
Page 20
Simulation of Fluorescent Concentrators | The Simulation
| August 10, 2008
Angular Experiment not fully specified → parameter variation: imperfect surface, scattering, properties of the blind, properties of the coupled cylinder Angular Distribution 0.018
experimental data simulation data 1 simulation data 2
0.016 0.014 0.012 light [fraction]
I
0.01 0.008 0.006 0.004 0.002 0 0
20
40
60
80 100 angle [degrees]
120
140
160
180
Page 21
Simulation of Fluorescent Concentrators | The Simulation
| August 10, 2008
Angular Experiment
Angular Distribution 0.016
experimental data simulation data
0.014
light [fraction]
0.012 0.01 0.008 0.006 0.004 0.002 0
0
20
40
60
80 100 angle [degrees]
120
140
160
180
Page 22
Simulation of Fluorescent Concentrators | The Simulation
| August 10, 2008
Fast Ray Tracing I
several millions of rays to be traced for one graph
I
exponentially more (in the number of parameters) needed for parameter variation
I
incoherent rays: ray packets/bundles not an option ray tracing kernel
I
I I I
triangle-based 4-ary BVH SAH-based construction
Page 23
Simulation of Fluorescent Concentrators | The Simulation
| August 10, 2008
Visuals I
direct comparison photograph/render
Page 24
Simulation of Fluorescent Concentrators | The Simulation
Non-Fluorescent Dragon
| August 10, 2008
Page 25
Simulation of Fluorescent Concentrators | The Simulation
Fluorescent Dragon
| August 10, 2008
Page 26
Simulation of Fluorescent Concentrators | Conclusion
| August 10, 2008
Future Work I
optimisation: I
concentrator stacks
I
photonic structures, mirrors, additional solar cells geometry optimisation different dyes (problem: infra-red range)
I I
Page 27
Simulation of Fluorescent Concentrators | Conclusion
| August 10, 2008
Summary I
fluorescent concentrators are made for improvement of efficiency and applicability of solar cells.
I
testing by simulation possible before (expensive) experimental testing
I
fast ray tracing makes automated parameter variation possible
I
general Monte Carlo simulation framework for physicists to gain insights in involved processes, for parameter optimisation (dye properties, geometry, ..) and for testing of new ideas
Questions?
Page 28
Simulation of Fluorescent Concentrators | Conclusion
Experimental Setup
| August 10, 2008
Simulation of Fluorescent Concentrators
Simulation of Fluorescent Concentrators | Fluorescent Concentrators
Page 2
| August 10, 2008
Solar Energy Research I
cooperation with the Fraunhofer Institute for Solar Energy Systems in Freiburg
I
raise efficiency – by concentration of light
I
problem: diffuse light
I
idea (around 1970): trap light inside medium
I
low efficiency → low research interest
I
interest increased again in the last years due to improved dyes, solar cells and concepts
Simulation of Fluorescent Concentrators | Fluorescent Concentrators
Page 3
| August 10, 2008
What is a Fluorescent Concentrator?
I
PMMA (acrylic glass) and fluorescent dye
I
concentrates direct and diffuse light on a solar cell principle:
I
I I I
absorption in dye re-emmitance according to Stokes shift (longer wavelength) total internal reflection (trap light inside medium)
Simulation of Fluorescent Concentrators | Fluorescent Concentrators
Page 4
| August 10, 2008
Absorption and Photoluminescence Spectra Absorption and Photoluminescence Spectra absorption photoluminescence
0.018 0.016
[dimensionless]
0.014 0.012 0.01 0.008 0.006 0.004 0.002 0 300
I
400
500 600 wavelength [nm]
re-absorption only possible in overlap
700
800
Simulation of Fluorescent Concentrators | Fluorescent Concentrators
Page 5
| August 10, 2008
Energy Loss Mechanisms
I
loss mechanisms have to be analysed for improvement
I
experimental testing of new concepts and analysis of physical processes difficult and expensive
I
analytical calculations and simulation complex and time-consuming
Simulation of Fluorescent Concentrators | The Simulation
Page 6
| August 10, 2008
Simulation Model ⇒ Monte Carlo (MC) method I
ray tracing of single photons (simplest simulation model, least error-prone)
I
uni-directional particle transport to avoid problems arising when connecting paths with different wavelengths
I
allows to investigate effects isolated from global path
Simulation of Fluorescent Concentrators | The Simulation
Page 7
| August 10, 2008
A Photon’s Path trough the Concentrator
I
sample spectrum AM 1.5 (MC inversion method)
I
sample start point
I
fix direction
Simulation of Fluorescent Concentrators | The Simulation
Page 8
| August 10, 2008
A Photon’s Path trough the Concentrator
I
calculate intersection with boundary
I
calculate reflection coefficient (Fresnel, Cauchy )
I
calculate new ray direction (Snell)
Simulation of Fluorescent Concentrators | The Simulation
Page 9
| August 10, 2008
A Photon’s Path trough the Concentrator
I
sample pathlength (MC inversion method, Lambert-Beer )
Page 10
Simulation of Fluorescent Concentrators | The Simulation
| August 10, 2008
A Photon’s Path trough the Concentrator
I
estimate absorption event
Page 11
Simulation of Fluorescent Concentrators | The Simulation
| August 10, 2008
A Photon’s Path trough the Concentrator
I
sample PL-spectrum (MC inversion method)
I
sample direction (MC inversion method)
I
sample pathlength (MC inversion method,) Lambert-Beer
Page 12
Simulation of Fluorescent Concentrators | The Simulation
| August 10, 2008
A Photon’s Path trough the Concentrator
I
calculate intersection with boundary
I
calculate reflection coefficient (Fresnel, Cauchy )
I
calculate new ray direction (reflection angle equals angle of incidence)
Page 13
Simulation of Fluorescent Concentrators | The Simulation
| August 10, 2008
A Photon’s Path trough the Concentrator
I
sample pathlength (MC inversion method, Lambert-Beer )
I
estimate absorption event
Page 14
Simulation of Fluorescent Concentrators | The Simulation
| August 10, 2008
A Photon’s Path trough the Concentrator
I
sample PL-spectrum (MC inversion method)
I
sample direction (MC inversion method )
I
sample pathlength (MC inversion method, Lambert-Beer )
I
calculate intersection with boundary
Page 15
Simulation of Fluorescent Concentrators | The Simulation
| August 10, 2008
A Photon’s Path trough the Concentrator
I
estimate absorption event → terminate ray
Page 16
Simulation of Fluorescent Concentrators | The Simulation
| August 10, 2008
Experiments and their Simulation I
physical measurement of input parameters: absorption spectrum, photoluminescence (PL-) spectrum, refractive indices and geometry
I
several experiments for analysis of fluorescent concentrators first step:
I
I I I
I
reproduce experiments “exactly” compare data to verify implementation explain small differences
second step: I I I
simulate experiments without limitations of “real” setup parameter variation evaluate results
Simulation of Fluorescent Concentrators | The Simulation
Page 17
| August 10, 2008
Reflection and Transmission Experiment
experiments with integrating sphere Transmission Spectrum
Reflection Spectrum 100 rays detected in reflection direction [%]
100
transmitted rays [%]
80
60
40
20
0 300
experimental data simulation data 400
500 600 Wavelength [nm]
700
800
experimental data simulation data
80
60
40
20
0 300
400
500 600 wavelength [nm]
700
800
Page 18
Simulation of Fluorescent Concentrators | The Simulation
| August 10, 2008
Absorption Absorption ≈ 1 − Reflection − Transmission calculated absorption as input → verification of simulated absorption Absorption Spectrum 100 rays with at least one absorption [%]
I I
experimental data simulation data
80
60
40
20
0 300
400
500 600 wavelength [nm]
700
800
Page 19
Simulation of Fluorescent Concentrators | The Simulation
| August 10, 2008
Angular Experiment Blind
PMMA half cylinder Light Photonic structure (optional) Detector Mirror (optional) Optical coupling
Page 20
Simulation of Fluorescent Concentrators | The Simulation
| August 10, 2008
Angular Experiment not fully specified → parameter variation: imperfect surface, scattering, properties of the blind, properties of the coupled cylinder Angular Distribution 0.018
experimental data simulation data 1 simulation data 2
0.016 0.014 0.012 light [fraction]
I
0.01 0.008 0.006 0.004 0.002 0 0
20
40
60
80 100 angle [degrees]
120
140
160
180
Page 21
Simulation of Fluorescent Concentrators | The Simulation
| August 10, 2008
Angular Experiment
Angular Distribution 0.016
experimental data simulation data
0.014
light [fraction]
0.012 0.01 0.008 0.006 0.004 0.002 0
0
20
40
60
80 100 angle [degrees]
120
140
160
180
Page 22
Simulation of Fluorescent Concentrators | The Simulation
| August 10, 2008
Fast Ray Tracing I
several millions of rays to be traced for one graph
I
exponentially more (in the number of parameters) needed for parameter variation
I
incoherent rays: ray packets/bundles not an option ray tracing kernel
I
I I I
triangle-based 4-ary BVH SAH-based construction
Page 23
Simulation of Fluorescent Concentrators | The Simulation
| August 10, 2008
Visuals I
direct comparison photograph/render
Page 24
Simulation of Fluorescent Concentrators | The Simulation
Non-Fluorescent Dragon
| August 10, 2008
Page 25
Simulation of Fluorescent Concentrators | The Simulation
Fluorescent Dragon
| August 10, 2008
Page 26
Simulation of Fluorescent Concentrators | Conclusion
| August 10, 2008
Future Work I
optimisation: I
concentrator stacks
I
photonic structures, mirrors, additional solar cells geometry optimisation different dyes (problem: infra-red range)
I I
Page 27
Simulation of Fluorescent Concentrators | Conclusion
| August 10, 2008
Summary I
fluorescent concentrators are made for improvement of efficiency and applicability of solar cells.
I
testing by simulation possible before (expensive) experimental testing
I
fast ray tracing makes automated parameter variation possible
I
general Monte Carlo simulation framework for physicists to gain insights in involved processes, for parameter optimisation (dye properties, geometry, ..) and for testing of new ideas
Questions?
Page 28
Simulation of Fluorescent Concentrators | Conclusion
Experimental Setup
| August 10, 2008
