### Joule-Brayton Process with Dry Air

- Using a more realistic medium:
- Dry Air = ideal gas with non-constant heat capacity c
_{p}(T)
- defined in MSL media library as DryAirNasa

- simply changing medium → model doesn't run
- inversion of h(T) with Dekker-Brent has wrong
start interval

- Problems with DryAirNasa:
- function for isentropic enthalpy uses constant c
_{p}
approximation
- provide better computation directly using
constant entropy

- compressor characteristics in ThermoDLR uses constant
c
_{p}
- inversion still gets wrong start interval
- problem: mass flow is zero initially
- use given mass flow as start value → model runs

- Finetuning:
- large deviations from expected results
- use procedure from Joule example above, including
simple test models
- internal inversions make model less stable

- fix: define ω instead of τ for compressor

- then closing circle and standard finetuning work