Annual Spectral Simulation

radiance-general
1

Hello everyone,

I am trying to find a way to simulate annual spectral lighting. As far as I know, this has not been done so far, and the link related to the explanation of this method in frads is inactive.

Two approaches come to mind:

  1. Performing a point-in-time simulation for each hour considered (which is very time-consuming).
  2. Using a 2-phase or 2.5-phase method (as far as I know, BSDF files do not contain spectral information, so higher-phase methods cannot be considered for now).

Since EPW files do not contain spectral lighting data, we need to compute spectral lighting values based on the EPW file in both approaches.

First Method:
We can use the gensky command to read DHI and DNI values from the EPW file and generate the spectral lighting for each hour. Then, we can use the rtrace command and repeat this process in a loop for all the required hours.
I did this for two days (the summer solstice and the winter solstice) for a room with a window facing south, assuming a completely clear sky, and I encountered two issues:

  1. Throughout the night, the result is never zero, and there is always about 1 lux in the results.
  2. At sunrise and sunset, the light level increases dramatically, to the extent that during the winter solstice, it returned an infinite amount.

Summer solstice

Column 1 Column 2 Column 3 Column 4 E F G H I J K L M N O P Q R S T U V W
hour lux mlux 760 740 720 700 680 660 640 620 600 580 560 540 520 500 480 460 440 420 400 380
0 0.9819072 1.653156 0.004737538 0.00519847 0.005616502 0.005784307 0.005934488 0.005696022 0.004997215 0.004341822 0.003660811 0.003646896 0.003928697 0.00529061 0.006477665 0.008452562 0.01015819 0.01122887 0.01092604 0.008732359 0.007998057 0.003737103
1 0.9932547 1.680741 0.004791212 0.005260693 0.005683729 0.005849821 0.006002401 0.005774796 0.005055576 0.004404763 0.003704285 0.003698601 0.003981432 0.005364812 0.006578475 0.008584659 0.01032523 0.01142015 0.01112108 0.008898703 0.008160373 0.003827536
2 0.9747171 1.653589 0.004739539 0.005196732 0.005619172 0.005776088 0.005928566 0.005690116 0.004983216 0.00434636 0.003651197 0.003653235 0.003922968 0.00528499 0.006474639 0.00845163 0.01016595 0.01122902 0.01093276 0.008741657 0.008004189 0.003754346
3 0.9466258 1.599335 0.004537025 0.004982268 0.005388448 0.005541242 0.005685906 0.00545832 0.004783568 0.004166448 0.003509401 0.003511418 0.00377911 0.005087696 0.006233878 0.008156768 0.009821578 0.01087646 0.01060548 0.008498393 0.00780484 0.003669302
4 2.00E+28 1.37E+28 1.16E+26 9.86E+25 8.35E+25 6.87E+25 6.62E+25 6.75E+25 8.52E+25 9.79E+25 1.19E+26 1.34E+26 1.39E+26 1.27E+26 9.41E+25 8.27E+25 4.23E+25 6.44E+25 9.40E+25 9.56E+25 7.74E+25 2.75E+25
5 1.34E+30 6.68E+29 9.82E+28 9.43E+28 9.03E+28 8.37E+28 7.28E+28 6.29E+28 5.22E+28 4.54E+28 3.94E+28 3.74E+28 3.28E+28 2.90E+28 2.45E+28 2.00E+28 1.46E+28 9.45E+27 7.42E+27 3.65E+27 3.65E+27 1.29E+28
6 753.5671 872.1439 2.214651 2.413981 2.619195 2.802916 3.031773 3.244238 3.33143 3.496228 3.558532 3.746145 3.929284 4.294053 4.561444 4.799912 5.069962 5.218303 4.886383 3.821409 3.439213 1.553192
7 1146.103 1350.239 3.294459 3.571918 3.858605 4.123303 4.469591 4.808788 4.989076 5.324719 5.522193 5.869709 6.185718 6.668661 7.063069 7.391369 7.860149 8.225389 7.89331 6.375254 5.963206 2.813145
8 1483.009 1720.557 4.224508 4.566031 4.916071 5.247862 5.673632 6.103543 6.352626 6.813684 7.105041 7.560411 7.961694 8.507759 8.979553 9.355426 9.948486 10.47423 10.15281 8.329597 7.941362 3.836226
9 1874.271 2157.834 5.402536 5.829337 6.257264 6.672217 7.197861 7.737888 8.061912 8.652125 9.040513 9.617952 10.11722 10.75005 11.3075 11.71972 12.43922 13.10284 12.74011 10.51552 10.02701 4.848726
10 2484.493 2843.367 8.00515 8.567589 9.142307 9.67007 10.36147 11.06896 11.46656 12.24914 12.73883 13.45685 14.02347 14.69767 15.24939 15.58758 16.33002 17.05279 16.53347 13.66341 12.64225 5.907139
11 3366.708 3554.91 10.98722 11.71116 12.41887 13.08376 13.95263 14.8033 15.29774 16.25723 16.84599 17.68529 18.30643 18.98765 19.50505 19.6683 20.37017 21.08556 20.32775 16.77241 15.00163 6.708296
12 3668.293 3937.402 12.58924 13.39938 14.1717 14.91217 15.87919 16.81767 17.35386 18.40073 19.0492 19.96467 20.61212 21.28996 21.77701 21.87281 22.57344 23.27555 22.4259 18.50803 16.29683 7.13532
13 3526.006 3712.897 11.43809 12.18793 12.9184 13.62177 14.52067 15.40776 15.92367 16.91685 17.54213 18.43168 19.07926 19.79164 20.33553 20.51924 21.26141 22.03576 21.27967 17.59925 15.73381 7.038675
14 2763.081 3023.901 8.509103 9.108073 9.710766 10.27897 11.02461 11.77643 12.19055 13.01752 13.55343 14.31475 14.91552 15.6099 16.20187 16.52781 17.30599 18.05904 17.5199 14.48318 13.35974 6.236944
15 2093.522 2376.287 6.181449 6.648799 7.127788 7.584684 8.163793 8.760904 9.106233 9.757858 10.18396 10.7986 11.3278 11.97501 12.54205 12.92863 13.67333 14.35559 13.93793 11.50224 10.89673 5.226104
16 1612.427 1874.438 4.550109 4.919816 5.304012 5.663653 6.117497 6.600801 6.870744 7.381204 7.708558 8.214532 8.654734 9.244506 9.761458 10.16991 10.82768 11.4169 11.09804 9.132917 8.742526 4.245687
17 1281.339 1488.437 3.613485 3.91236 4.220418 4.507505 4.87683 5.248502 5.450979 5.83237 6.058391 6.442307 6.782934 7.289102 7.713238 8.048959 8.557195 8.979622 8.642918 7.022586 6.612406 3.150304
18 905.2373 1049.202 2.614088 2.841863 3.078401 3.298181 3.572769 3.837283 3.962346 4.191853 4.301172 4.549923 4.788761 5.207121 5.527508 5.804271 6.153256 6.381895 6.038663 4.787044 4.381734 2.017045
19 4.55E+34 3.28E+34 3.78E+32 3.72E+32 3.49E+32 3.30E+32 3.02E+32 2.90E+32 2.81E+32 2.91E+32 2.91E+32 2.93E+32 2.94E+32 2.88E+32 2.59E+32 1.95E+32 1.80E+32 9.90E+31 1.82E+32 1.80E+32 1.45E+32 5.64E+31
20 1.52E+36 1.22E+36 9.23E+33 8.53E+33 7.22E+33 6.07E+33 4.96E+33 4.71E+33 4.76E+33 5.91E+33 6.77E+33 8.22E+33 9.37E+33 9.98E+33 9.46E+33 7.46E+33 6.99E+33 3.83E+33 4.87E+33 9.13E+33 9.28E+33 8.38E+33
21 0.9470851 1.588152 0.004553751 0.004995359 0.005406708 0.005550637 0.005693509 0.00546094 0.004788497 0.004166326 0.003507859 0.003508733 0.003764045 0.005070028 0.006205697 0.008110294 0.009760167 0.01078723 0.0105149 0.008414212 0.007719036 0.00362389
22 0.9651915 1.615223 0.004669797 0.005127898 0.005548396 0.005705163 0.005852715 0.005622499 0.004922985 0.004292324 0.003609816 0.003608464 0.003878686 0.005220316 0.006402039 0.008366104 0.0100617 0.01112372 0.01083805 0.008667163 0.007935052 0.003720627
23 0.9503759 1.602313 0.004678104 0.005129392 0.005549266 0.005713807 0.005864501 0.005633109 0.004932548 0.004308036 0.003620218 0.003622791 0.003889884 0.005249696 0.006436735 0.008405518 0.01011703 0.01118552 0.0109017 0.008714384 0.007985275 0.003740322

Second Method:
I’ve drawn a work flow for this method

ASPD
ASPD1182Ă—683 27.4 KB

This method requires a command similar to gendaymtx to extract all annual data (or selected hours) from the EPW file, compute their spectral values, and return them as a three-dimensional matrix consisting of time, sky patches, and spectral data. Marshal Maskarenj has done something similar with EPW2AnnualSpectra, which you can find here (although I couldn’t find the source code for the executable and don’t know the exact method used). This process takes about 20 minutes per EPW file.
Afterward, due to variations in material reflectance and transmittance, we need to generate a DC matrix for each spectral channel (another 3D matrix). Finally, these matrices are multiplied together in sequence, resulting in a 3D matrix of sensors, time, and spectrum. By applying relevant response curves, such as the ipRGC cell sensitivity curve, we can obtain the final values.
Untitled-1
Untitled-13508Ă—1938 530 KB

Can we add these capabilities to commands, or do we need something completely new?

I also had a few requests:

  1. The output of the rtrace command seems to go from 780 to 380. Can you change the output direction from 380 to 780? In the pyradiance library, the output is also in string format, and it would be better if it could be converted to a list.
  2. The rtpict command does not work on Windows when spectral settings are enabled. It returns an empty hsr file, while the same python code works correctly in the Linux.

Thanks in advance

2

I can’t give you a complete answer, but I can point you to a few resources and ideas. @Taoning_Wang and I laid out some of the processes for annual spectral simulations in our tutorials at last year’s Radiance workshop, whose presentations you may find linked here.

Some additional pointers:

  1. Although you can think of a spectral matrix as a 3-D matrix, it is really a stack of matrices when you look at the operations rather than a tensor. This data type is well-supported in rmtxop and the new rcomb tool.

  2. You can use the -c option of rmtxop to transform the RGB color space of a BTDF file so it may be used in a spectral operation like so:

rmtxop spec_samp.mtx -c spec_samp.mtx windowBTDF.xml skycoef.mtx > combDC.mtx

where the first matrix is computed using rfluxmtx or rcontrib with the desired spectral sampling, and this matches the sky coefficient matrix given last. The “-c spec_samp.mtx” converts “windowBTDF.xml” to the same spectral sampling. The combined daylight coefficient matrix then corresponds to the given window BTDF, implementing the 3-phase method.

  1. It is not advised to change the spectral ordering in the file, as going from long to short wavelengths corresponds to the standard RGB and XYZ ordering of tristimulus values.

The new gensdaymtx tool can output spectral sky matrices directly – did you try it? @Taoning_Wang is working to add direct EPW input, but that’s not quite ready, yet.

I hope this is helpful.
Cheers,
-Greg

1 Like