Possibly incorrect assignment of RADCAL_SPECIES2KAPPA for Wide Band Model (NUMBER_SPECTRAL_BANDS > 1)

[mst-parkhachev]

Dear developers,

As noted in #14799, the calculation results when using the Wide Band Model depend on the PATH_LENGTH.
Appendix C in the Tech. Ref. Guide states that when the number of bands is more than one, only the Planck mean coefficient, which is independent of the PATH_LENGTH, is used in calculating the absorption coefficient. As far as I understand, the Wide Band Model implies the case of N==6.

I took a quick look at the implementation of the absorption coefficient calculation in radi.f90.

      CALL SUB_RADCAL(AMEAN,AP0,RADIANCE,TRANSMISSIVITY)
      IF (NSB==1 .AND. PATH_LENGTH > 0.0_EB) THEN
         RADCAL_SPECIES2KAPPA(NS,J,K,1) = MIN(AMEAN,AP0)
      ELSE
         RADCAL_SPECIES2KAPPA(NS,J,K,IBND) = AMEAN/BBF
      ENDIF

As far as I understand, AMEAN is EFFECTIVE_ABSORPTION (or path mean), not PLANCK_MEAN_ABSORPTION.
This doesn't seem to match what is written in the manual for the NSB > 1 case.

Quote:

[shostikk]

This line was changed in last January but documentation was not updated. My mistake to forget it - sorry!
#14068

A few weeks ago, we realized this change may be only accurate for non-sooting mixtures. For soot-containing mixtures, RADCAL adds the tails of the soot Planck function to the band integral of transmissivity. Problem is now it adds it to every band, not just first and last. I think I know how to fix.

[mst-parkhachev]

Do I understand correctly that in the new algorithm the result of calculating the heat flux in a Wide Band Model depends on the parameter PATH_LENGTH? That is, to estimate the magnitude of the heat flux at different distances from the fire source, it is necessary to run a simulation with different PATH_LENGTH values.

[shostikk]

Yes, it does, in the similar manner to the gray calculations. We think the Planck mean - even for the limited spectral bands - can only work for pure emission calculation.

Need to run simulations with different PATH_LENGTH values depends... if there was just an emitting flame in transparent atmosphere, I think you should only consider the path length of the emitter. But if the atmosphere is absorbing, and you try to capture that, then the path length should correpond to a relevant attenuation distance.

[shostikk]

Code changed now so that absorption coefficient is smaller of the effective and Planck mean values, whenever path length is greater than zero. For zero path lenght, we use Planck mean. The logic is same for single band and multiple bands.

Tech ref guide updated.
See #15067