[PHYSICS BUG][MEDIUM] Galactic confusion noise absent from LISA PSD

[JasperSeehofer]

PSD implementation: master_thesis_code/LISA_configuration.py
Unused constants: master_thesis_code/constants.py:77–83

power_spectral_density_a_channel() implements only instrumental (OMS + test-mass) noise. The galactic confusion-noise parameters (LISA_PSD_A, LISA_PSD_ALPHA, LISA_PSD_F2, LISA_PSD_BETA, LISA_PSD_KAPPA, LISA_PSD_GAMMA, LISA_PSD_FK) are defined in constants.py but never added to the PSD.

Galactic confusion noise dominates the LISA sensitivity from ~0.1 to ~3 mHz — exactly where many EMRI signals peak. Without it, computed SNRs and Cramér–Rao bounds are systematically too optimistic.

The model (Babak et al. 2023, Eq. 17):
179770S_\mathrm{gal}(f) = A,f^{-7/3}\exp!\left(-f^\alpha+\beta f\cos(\kappa f)\right)\frac{1+\tanh!\left(\gamma(f_k-f)\right)}{2}179770
with T_obs-dependent parameters for 0.5/1/2/4-year observations.

Fix: Implement galactic_confusion_noise(frequencies, T_obs) and add it to the total PSD in power_spectral_density_a_channel().

Reference: Babak et al. (2023), arXiv:2303.15929, Eq. (17) and Table 1.

Physics Change Protocol required.