README.md

2026SGSH

Codex/ChatGPT (July 2026)

Summary

Using consistently analyzed H-mode measurements from 111 Alcator C-Mod, 43 ASDEX Upgrade, and 45 JET discharges, the study models the electron separatrix density $n_{e,\mathrm{sep}}$ from engineering parameters. A log-link regression finds positive dependence on divertor neutral pressure and $P_{\mathrm{SOL}}/R_{\mathrm{geo}}$, negative dependence on toroidal field and machine size, and negligible plasma-current dependence, achieving $R^2=0.91$ and 19% normalized RMSE. An independently derived two-point-model scaling coupled to divertor geometry gives similar device constants and parameter trends. The resulting hybrid formula retains physics-based normalization and regression exponents, predicts the three-machine database within a factor of 1.5, and gives projections for ITER, SPARC, DTT, JT-60SA, and COMPASS-U consistent with available SOLPS results. Applicability is presently limited mainly to similar, mostly unseeded closed-divertor H-mode conditions.

Contributions

1. Built a cross-machine H-mode separatrix-density database using a consistent analysis procedure.
2. Quantified engineering-parameter dependencies and device-specific constants with generalized linear regression.
3. Derived an engineering-input two-point-model expression including divertor and magnetic geometry.
4. Combined physics-based normalization with empirical exponents into a formula accurate within a factor of 1.5.
5. Projected separatrix densities for five next-step devices and checked them against available SOLPS simulations.