DMGB Systems - Second-law boundary audit and reproducible MHD ledger
DMGB | Independent researcher | Working note v3 | 5 October 2026

START HERE
publication/manuscript.pdf contains the methods and evidence note.
publication/efficiency-note.pdf distinguishes kinetic and isentropic metrics.
Neither document is an arXiv submission or a peer-reviewed publication.

REPRODUCE
Python 3.10 or later. Install numpy, scipy and matplotlib for the complete run:
  python -m pip install numpy scipy matplotlib
  python threshold_model.py --no-plots
  python dimensionless.py
  python inlet_reduced.py
  python second_law.py
  python reproduce_shneider_source_check.py
  python -m unittest discover -s . -p "test_*.py" -v
  python independent_validation.py
  python test_inlet_review.py
  python calibration-check.py
To regenerate PDFs, install reportlab, then run python build_pdfs.py.
The original threshold and exact-rational checks use only the standard library.

WHAT THE MAP MEANS
x=e/d; alpha=a/d; chi=c/d; r=g*h*s. The requested x-versus-r map
still requires generator conversion g and auxiliary fraction alpha.
With z=(e+a)/(g*d), chi_required=1-r*(1-z). This universal accounting
map contains no chosen dimensional extraction value. Colors are requirements,
not demonstrated benefits. Electrical feasibility requires z<=1 and compatible
components g>=r. Zero extraction cannot be normalized by d.
Reference r=0.36 and z=0.05 requires chi=0.658. The historical d=400 kJ/kg
slice therefore requires 263.2 kJ/kg. It is retained only for continuity.

REDUCED INLET
inlet_reduced.py models a constant-area perfect-gas scalar Faraday generator,
then a normal shock and ideal expansion at matched back pressure. It has
no combustion, plasma kinetics or installed thrust calculation. Local inlet
Mach is not automatically flight Mach. Sigma and B enter as sigma*B**2.
Gas state, length, loading and back pressure must also be specified.
Default d=7.784959 kJ/kg; c_equiv=4.473258 kJ/kg; chi=0.574603.
Free ionization and auxiliaries with h=0.60 give a negative proxy:
Delta=-0.509116 kJ/kg. This result is conditional, not an engine limit.
The 27-case sweep has one sonic-guard interruption and 26 completed lengths.
Interrupted cases have no fabricated final credit.

CHECKS AND LIMITS
39 unittest methods include exact dimensionless identities, scaling, domains,
independent RK4 integration, shock conservation and invalid-input checks.
The original checks also cover 10,000 rational cases, 144 boundary cases
and three root cases. Arithmetic and solver checks do not calibrate physics.
Optional --project-root arguments in independent_validation.py and
calibration-check.py compare the historical 81 slices or source-PDF hashes.
Standalone source arithmetic is not presented as new source retrieval.

PRIOR ART AND PUBLICATION
prior-art-v3.txt/json update the retained v2 literature screen, overlap and access.
shneider-validation-v3.txt/json contain the external benchmark assessment.
Four journal full texts remain unresolved; novelty is not established.
The manuscript uses DMGB, Independent researcher. Identity no longer blocks
research. Final author review, account, license and applicable endorsement
remain necessary for actual submission. No submission or outreach occurred.

SOURCE AND RIGHTS BOUNDARY
Original plots come from the documented equations. Scientific publications
are cited, not redistributed. No third-party PDFs, page screenshots, private
outreach, credentials or authentication configuration are included.
No distribution license is assigned on the author's behalf.
manifest.json records SHA256 hashes. Rebuilding plots/PDFs may change
metadata bytes; compare numerical outputs separately.

V3 SECOND-LAW RESULT
reference_loss gives T_env*R*ln(1/Pi) for a passive no-work adiabatic
uniform/equivalent perfect-gas reference. It does NOT cap kinetic c_equiv.
At 70% recovery and 220K, destroyed exergy is 22.52438 kJ/kg, while
ideal-nozzle kinetic loss is 23.71208 kJ/kg. Residual thermal exergy
accounts for the difference. Read the five-page manuscript for the exact
compatible bound, conditional exergy screening and double-counting caution.
Shneider2004 no-MHD averaged recoveries give equivalent diagnostics
9.89284/72.75078/115.33589 kJ/kg at Mach5/8/10 and a220K dead state.
These are source-anchored calculations, not mass-flux entropy measurements.
The current local inlet model remains NOT VALIDATED against that complete
published inlet. Source arithmetic and independent solver tests are separate.
Gamma_X and beta_save,crit are NEW conditional exergy screening labels.
B_controllable is reserved, not assigned a numerical or physical control limit.
methods-supplement-v2.txt preserves the preceding detailed working methods;
v3 corrections take precedence. No correspondence or sender address is included.
