Radiation model update in StFlow.cpp #965

lavrenyukiv · 2021-01-25T18:06:44Z

Changes proposed in this pull request

Absorption coefficients for CO2, H2O, CO, CH4, OH molecules are added,
Spectrums from https://hitran.org/ are used for evaluated coefficients,
Polynomials are changed by linear interpolation in logarithmic coordinates between tabulated values,

index error fixed.

error fixed.

codecov · 2021-01-25T19:24:20Z

Codecov Report

Merging #965 (5db79c0) into main (e49d62b) will decrease coverage by 0.02%.
The diff coverage is 55.00%.

@@            Coverage Diff             @@
##             main     #965      +/-   ##
==========================================
- Coverage   71.26%   71.23%   -0.03%     
==========================================
  Files         377      377              
  Lines       46299    46371      +72     
==========================================
+ Hits        32993    33031      +38     
- Misses      13306    13340      +34

Impacted Files	Coverage Δ
src/oneD/StFlow.cpp	`86.24% <55.00%> (-4.42%)`	⬇️

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update e49d62b...89cc361. Read the comment docs.

ischoegl

Thanks for working on this. While others may have more comments, I have one observation to share.

ischoegl · 2021-01-26T00:05:31Z

src/oneD/StFlow.cpp

@@ -94,9 +94,12 @@ StFlow::StFlow(ThermoPhase* ph, size_t nsp, size_t points) :
    setupGrid(m_points, gr.data());

    // Find indices for radiating species
-    m_kRadiating.resize(2, npos);
+    m_kRadiating.resize(5, npos);


There has been some work to make the identification of species more flexible (which would mean that lowercase species names would be recognized). Rather than hard coding names in the subsequent lines, specifying the elemental composition will in most cases work (there usually is only one isomer). The function name is findIsomers

I changed this part.

It would be great if the user could select the components involved in the radiation himself.

It would probably be better to load the absorption coefficient data from a separate file, so that the user can add his own dependencies.

Unfortunately, my experience with С++ is not enough.

I agree that it would be great to select species / import parameters instead of hard coding them. There are likely multiple ways of getting this done. I can probably give you pointers, but it’s up to @bryanwweber and @speth ...

bryanwweber

Thank you for contributing these changes! It's great to see work ongoing with the radiation models. Other than my style comment below, and the comment from @ischoegl, I have three other suggestions to help improve this:

Please keep line lengths to less than 88 characters
Please include the citation for where this data is derived from. Is there a specific license associated with the HITRAN data? I couldn't determine from a brief look at their website.
Please add a test for the added code and/or an example that shows the effect of radiation for these species

Thanks again for your contribution! If there's anything I can do to help you with these comments, please let me know!

bryanwweber · 2021-01-26T01:30:06Z

src/oneD/StFlow.cpp

-
+        // Temperatures for Planck optical path length evaluation, K
+        const int OPL_table_size = 28;
+        const doublereal TemperatureOPL[OPL_table_size] = {300.0,400.0,500.0,600.0,700.0,800.0,900.0,1000.0,1100.0,1200.0,1300.0,1400.0,1500.0,1600.0,1700.0,1800.0,1900.0,2000.0,2100.0,2200.0,2300.0,2400.0,2500.0,2600.0,2700.0,2800.0,2900.0,3000.0};


Please don't use doublereal in new code, you can just use double. I also think this can be replaced with a std::vector. Finally, please indent your code with spaces, instead of tab characters.

Done

I downloaded HAPI library from https://hitran.org/hapi/ (citation: R.V. Kochanov, I.E. Gordon, L.S. Rothman, P. Wcislo, C. Hill, J.S. Wilzewski, HITRAN Application Programming Interface (HAPI): A comprehensive approach to working with spectroscopic data, J. Quant. Spectrosc. Radiat. Transfer 177, 15-30 (2016), https://doi.org/10.1016/j.jqsrt.2016.03.005).
Then I downloaded spectra for more important for combustion tasks molecules. It seems that the number of spectra available for download is limited (or I haven't figured it out yet). Then I computed spectrum averaged coefficients for optically thin medium (with Planck function) by myself. I visually compared my results with the results presented earlier(http://www.sandia.gov/TNF/radiation.html), but I am not completely sure.

I prepared example *.py file. In which place it could be placed?

- replace tabs with spaces - replace doublereal types with double

ischoegl

One more comment beyond my initial observation.

ischoegl · 2021-01-27T23:58:46Z

src/oneD/StFlow.cpp

-                for (size_t n = 0; n <= 5; n++) {
-                    k_P_H2O += c_H2O[n] * pow(1000 / T(x, j), (double) n);
+                for (int k = 0; k < OPL_table_size; k++) {
+                    if (T(x, j) < TemperatureOPL[k]) {


This implementation could be improved in terms of computational efficiency as it appears that the same temperature lookup loop is called for each of the species. Beyond, It should be possible to consolidate the species-specific code sections, which - as far as I can tell - are mostly identical except for the species index.

PS: I am also not convinced that a linear interpolation within a lookup table is an improvement over a polynomial fit - unless the approximation is poor for the newly added curves. (Btw, how do new H20/CO2 curves compare to the previous implementation?)

I will make a comparison between old and new curves for H2O and CO2.
Linear interpolation is not an improvement, but rather a more general assumption. For other species we have not polynomial type assumption and polynomial coefficients. But we have spectra for it. Integration of spectra is a very expensive operation in terms of calculations. So tabulated data is accurate and fast.
The use of a polynomial can lead to negative values of the absorption coefficients.
I found that CO, CH4 and OH molecules spectra integrated before and presented by me are wrong, so I will change this data.

Plots in logarithmic coordinates are at the left part, the same plots in linear coordinates are at the right part.
Results for H2O look similar. Most likely the polynomial approximation was done on a linear scale. So there is a large difference in the logarithmic scale at high temperatures. So linear interpolation in logarithmic scale looks preferable.
The results for CO2 are very different. I will try to find the data from the original source.

If we compute S * a * (T^4-300^4), where S is Stefan - Boltzmann constant, we can make an estimate of the absorbed energy(W/m3/atm). Here the differences between approximations will be huge.

@lavrenyukiv ... thank you - this looks interesting. I was not involved in the original model, so it's hard to tell. I agree that the differences for energy are substantial. The original polynomial coefficients appear to be taken directly from the TNF link you had posted (where it states that "These curve fits were generated for temperatures between 300K and 2500K and may be very inaccurate outside this range."; the fit itself is for 1000/T so deviations may not look as dramatic?). I am not familiar with RADCAL and assumptions used therein, so some legitimate differences may exist.

PS: There appears to be an updated version of the radcal code on GitHub.

PPS: Incidentally, there was a post on the User Group about unphysically low temperatures but it's hard to tell whether these are related as no code was posted.

Yes, it is a negative absorption coefficient at low temperatures.

I am trying to compare spectra.
Results for CO2 from https://github.com/firemodels/radcal/blob/master/Documentation/Figures/CO2_300K.png

and spectra from hapi

full range

the RADCAL range

detailed view

Spectra look similar in the specified wavenumber range. The hapi has a greater resolution.

- Added coefficients for CO2, H2O, CO, CH4, OH, N2O, NH3, H2O2, H2CO, NO2, NO, HCN, C2H2, C2H4, C2H6, C2N2, C4H2, HCOOH, HNO3, O3 - Temprature range expanded to 200 - 3500 K

- Definitions for radiation model added

ischoegl · 2021-02-09T18:40:03Z

@lavrenyukiv ... I just noticed that there may be some related work on radiation properties posted in enhancements (Cantera/enhancements/issues/72). I believe it mainly addresses properties and not the implementation itself.

lavrenyukiv · 2021-02-12T07:05:02Z

Cross validation results for HF, HCl, HBr, CO.
Points are from the article DOI: 10.1115/1.1416689

speth · 2021-02-19T21:05:55Z

I think some improvements and generalizations to the way radiation is handled in Cantera's 1D code would be very welcome.

With the interest in extending the set of species handled by the radiation model, and the fact that data for so many more species are available from HITRAN, I think it would make sense to start reading this data in from a YAML input file, so users could provide their own data for additional species. The data could either be include as a new section within each species definition, or in a separate file. I'd also like to make sure that we keep the door open to allowing other radiation models, like the ones being suggested in Cantera/enhancements#72.

I think the choice of log-linear interpolation seems like a reasonable alternative to using polynomial fits. It's certainly familiar enough from its use in the PLOG rate constant calculations.

ischoegl · 2021-10-29T02:27:26Z

@lavrenyukiv … while some time has passed since this was initially discussed, I believe there have been improvements in the YAML interface that should make the import of external data a lot simpler, especially if you interact from Python. Within C++, things are handled using an object named AnyMap that is fairly easy to use, and should work nicely for a more generic approach to radiative heat losses.

Stale

ischoegl · 2022-06-08T01:22:50Z

@lavrenyukiv ... are you still pursuing this PR or should it be closed?

ischoegl · 2024-11-05T02:24:48Z

Superseded by #1799

lavrenyukiv added 3 commits January 25, 2021 22:36

Update StFlow.cpp

05c7925

Update StFlow.cpp

6201d6a

index error fixed.

Update StFlow.cpp

234fa4b

error fixed.

lavrenyukiv changed the title ~~Update StFlow.cpp~~ Radiation model update in StFlow.cpp Jan 25, 2021

ischoegl reviewed Jan 26, 2021

View reviewed changes

bryanwweber previously requested changes Jan 26, 2021

View reviewed changes

Update StFlow.cpp

5db79c0

- replace tabs with spaces - replace doublereal types with double

ischoegl requested changes Jan 27, 2021

View reviewed changes

lavrenyukiv added 3 commits February 2, 2021 03:15

Update StFlow.cpp

0cfb26f

- Added coefficients for CO2, H2O, CO, CH4, OH, N2O, NH3, H2O2, H2CO, NO2, NO, HCN, C2H2, C2H4, C2H6, C2N2, C4H2, HCOOH, HNO3, O3 - Temprature range expanded to 200 - 3500 K

Update StFlow.h

ea5d074

- Definitions for radiation model added

Update StFlow.cpp

89cc361

lavrenyukiv requested a review from bryanwweber February 1, 2021 23:08

ischoegl mentioned this pull request Feb 9, 2021

Radiation property library Cantera/enhancements#72

Open

bryanwweber removed their request for review February 10, 2022 20:48

ischoegl added the 1D Flames label Aug 6, 2022

wandadars mentioned this pull request Oct 22, 2024

Revive 1D radiation improvements of PR965 #1799

Open

ischoegl closed this Nov 5, 2024

ischoegl mentioned this pull request Dec 28, 2024

Inclusion of alternative radiation models and species Cantera/enhancements#218

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Radiation model update in StFlow.cpp #965

Radiation model update in StFlow.cpp #965

lavrenyukiv commented Jan 25, 2021 •

edited

Loading

codecov bot commented Jan 25, 2021 •

edited

Loading

ischoegl left a comment •

edited

Loading

ischoegl Jan 26, 2021 •

edited

Loading

lavrenyukiv Feb 3, 2021

ischoegl Feb 3, 2021

bryanwweber left a comment

bryanwweber Jan 26, 2021

lavrenyukiv Jan 26, 2021 •

edited

Loading

ischoegl left a comment

ischoegl Jan 27, 2021

ischoegl Jan 28, 2021 •

edited

Loading

lavrenyukiv Jan 28, 2021 •

edited

Loading

lavrenyukiv Jan 28, 2021 •

edited

Loading

ischoegl Jan 28, 2021 •

edited

Loading

lavrenyukiv Jan 28, 2021

lavrenyukiv Jan 30, 2021 •

edited

Loading

ischoegl commented Feb 9, 2021

lavrenyukiv commented Feb 12, 2021

speth commented Feb 19, 2021

ischoegl commented Oct 29, 2021

ischoegl commented Jun 8, 2022

ischoegl commented Nov 5, 2024

Radiation model update in StFlow.cpp #965

Radiation model update in StFlow.cpp #965

Conversation

lavrenyukiv commented Jan 25, 2021 • edited Loading

codecov bot commented Jan 25, 2021 • edited Loading

Codecov Report

ischoegl left a comment • edited Loading

Choose a reason for hiding this comment

ischoegl Jan 26, 2021 • edited Loading

Choose a reason for hiding this comment

lavrenyukiv Feb 3, 2021

Choose a reason for hiding this comment

ischoegl Feb 3, 2021

Choose a reason for hiding this comment

bryanwweber left a comment

Choose a reason for hiding this comment

bryanwweber Jan 26, 2021

Choose a reason for hiding this comment

lavrenyukiv Jan 26, 2021 • edited Loading

Choose a reason for hiding this comment

ischoegl left a comment

Choose a reason for hiding this comment

ischoegl Jan 27, 2021

Choose a reason for hiding this comment

ischoegl Jan 28, 2021 • edited Loading

Choose a reason for hiding this comment

lavrenyukiv Jan 28, 2021 • edited Loading

Choose a reason for hiding this comment

lavrenyukiv Jan 28, 2021 • edited Loading

Choose a reason for hiding this comment

ischoegl Jan 28, 2021 • edited Loading

Choose a reason for hiding this comment

lavrenyukiv Jan 28, 2021

Choose a reason for hiding this comment

lavrenyukiv Jan 30, 2021 • edited Loading

Choose a reason for hiding this comment

ischoegl commented Feb 9, 2021

lavrenyukiv commented Feb 12, 2021

speth commented Feb 19, 2021

ischoegl commented Oct 29, 2021

ischoegl commented Jun 8, 2022

ischoegl commented Nov 5, 2024

lavrenyukiv commented Jan 25, 2021 •

edited

Loading

codecov bot commented Jan 25, 2021 •

edited

Loading

ischoegl left a comment •

edited

Loading

ischoegl Jan 26, 2021 •

edited

Loading

lavrenyukiv Jan 26, 2021 •

edited

Loading

ischoegl Jan 28, 2021 •

edited

Loading

lavrenyukiv Jan 28, 2021 •

edited

Loading

lavrenyukiv Jan 28, 2021 •

edited

Loading

ischoegl Jan 28, 2021 •

edited

Loading

lavrenyukiv Jan 30, 2021 •

edited

Loading