Opened 2 years ago

Last modified 2 years ago

#332 new defect - wrong answer

default zone thickness gives incorrect line emissivity

Reported by: Gary J. Ferland Owned by: nobody
Priority: blocker Milestone: C17_branch
Component: radiative transfer Version: trunk
Keywords: Cc:

Description

I was contacted by a participant at the last Belfast workshop who was comparing the workshop version of Cloudy with c13. His group was generating large grids of line emissivities for various densities and kinetic temperatures, for gas exposed to the H&M background. He found very large differences in the emissivities of strong lines such as Lya. The problem can be distilled to the attached sims.

The sim is a single zone, constant temperature, exposed to the H&M background. I compared the trunk with c13_branch. The attached tests report the emissivity of Lya in the fiile *.em.

The file sdr.in behaves correctly and gets the same emissivity with c13 and trunk. It sets the zone thickness to 1 cm.

c13dr.in uses c13 and lets the thickness of the zone be set by the code. tdr.in does the same thing for the trunk. Here are the results for the zone thickness and the emissivity, erg/cm3/s in Lya

sim thickness 4pi J (Lya)
sdr 1 -28.7937
c13dr 4.647E+20 -29.0139
tdr 4.613E+20 -35.9298

The emissivity is per unit volume so should be the same for all three, and equal to the first. Both c13 and the trunk get the same answer for sdr.

Attachments (3)

sdr.in (329 bytes) - added by Gary J. Ferland 2 years ago.
c13dr.in (322 bytes) - added by Gary J. Ferland 2 years ago.
tdr.in (332 bytes) - added by Gary J. Ferland 2 years ago.

Download all attachments as: .zip

Change History (4)

Changed 2 years ago by Gary J. Ferland

Attachment: sdr.in added

Changed 2 years ago by Gary J. Ferland

Attachment: c13dr.in added

Changed 2 years ago by Gary J. Ferland

Attachment: tdr.in added

comment:1 Changed 2 years ago by rjrw

The value is set by xInten at lines_service.cpp:255. It seems likely that 100pc of material will have an effect on that. Running a spatially-resolved calculation on the trunk seems to confirm that, giving a face value comparable to the 1cm run on c13, and an average comparable to the 1 zone, 100pc c13 run.

The far lower value for the 1 zone calculation on the trunk is troubling, particularly since the spatially-resolved calculation is so consistent with c13. Disabling corr_xIntensity() at lines_service.cpp:869, the reported emissivity for the 100pc trunk calculation becomes -28.8564, pretty consistent with c13 -- and with the spatially-resolved calculation on the trunk.

Note: See TracTickets for help on using tickets.