Using EOS module in Pressure-Temperature mode

[Quacken8]

Hello,

I'm writing a thesis on solar structure based on one particular article (Schüsser & Rempel 2005) and I'd like to reproduce their results. It would be very useful for me to use MESA EOS in PT mode as that simplifies many of my caluclations significantly.
In the sample_eos.f90 there's this one part that seems to suggest that MESA is able to run in PT mode:

! call the density-temperature based eos

! composition derivatives by isotope

         call eosDT_get( &
               handle, species, chem_id, net_iso, xa, &
               Rho, log10(Rho), T, log10(T), &
               res, d_dlnd, d_dlnT, d_dxa, ierr)


! report the results

         call PrettyOut(Rho, T, abar, zbar, res, d_dlnd, d_dlnT)


! example calling the pressure-temperature based eos
!         Pgas = exp(res(i_lnPgas))
!         call eosPT_get( &
!               handle, Z, X, abar, zbar, &
!               species, chem_id, net_iso, xa, &
!               Pgas, log10(Pgas), T, log10(T), &
!               Rho, log10Rho, dlnRho_dlnPgas_const_T, dlnRho_dlnT_const_Pgas, &
!               res, d_dlnd, d_dlnT, d_dabar, d_dzbar, ierr)

however it seems like this way was abandoned and I'm not savvy enough in fortran or thermodynamics yet to revive it. Can you help me make this module run in PT mode?

[rjfarmer]

eosPT_get still exists in the code

mesa/eos/public/eos_lib.f90

Line 259 in bb123bb

subroutine eosPT_get( &

you just need to uncomment the code in the sample_eos.f90 file (and check that the order of the arguments still match)

[Quacken8]

Oh, cool! So the results it gives are consistent in terms of physics?

[rjfarmer]

Yes, under the hood eosPT is just eosDT but we handle the transformation between inputs/outputs for you.

mesa/eos/private/eospt_eval.f90

Line 154 in 80d1e85

subroutine Get_PT_Results_using_DT( &

[Quacken8]

Amazing, thank you!

[Quacken8]

Follow up question: eos DT returns derivatives alongsite the resutls, eos_def names them d_dlnd and d_dlnT. When I run eos in PT mode, do these same arrays now hold (∂/∂lnP)_T and (∂/∂lnT)_P respectively?

[evbauer]

Not quite, but the routines should return enough information for you to construct what you need in terms of derivatives. In the code block where the public eosPT_get routine is declared, we can look through the variable declarations for the arguments to that routine to find the output variables:

mesa/eos/public/eos_lib.f90

Lines 292 to 307 in bb123bb

	! OUTPUT
	real(dp), intent(out) :: Rho, log10Rho ! density
	real(dp), intent(out) :: dlnRho_dlnPgas_const_T
	real(dp), intent(out) :: dlnRho_dlnT_const_Pgas
	real(dp), intent(inout) :: res(:) ! (num_eos_basic_results)
	! partial derivatives of the basic results
	real(dp), intent(inout) :: d_dlnRho_const_T(:) ! (num_eos_basic_results)
	! d_dlnRho_const_T(i) = d(res(i))/dlnd|T,X where X = composition
	real(dp), intent(inout) :: d_dlnT_const_Rho(:) ! (num_eos_basic_results)
	! d_dlnT_const_Rho(i) = d(res(i))/dlnT|Rho,X where X = composition
	real(dp), intent(inout) :: d_dxa_const_TRho(:,:) ! (num_eos_d_dxa_results, species)
	! d_dxa_const_TRho(i) = d(res(i))/X|T,Rho,X where X = composition

This shows that you still have the usual vectors of derivatives such as d_dlnRho_const_T, and you also have a couple of additional derivative quantities dlnRho_dlnPgas_const_T and dlnRho_dlnT_const_Pgas, which you can use to construct (∂/∂lnP)_T and (∂/∂lnT)_P via the chain rule.