[grmtrkngtn]

Description

Add a small lower bound to the electron-density profile before it is used in the Sauter bootstrap-current calculation:

Calculate electron and ion density profiles

ne = np.maximum(ne, 1.0e-6)

context

During normal PROCESS optimisation the density profile remains within the expected physical domain, but UQ sampling deliberately explores a much wider region of parameter space, including trial points that may be far from a feasible solution.

At some of these trial points, one or more values in the calculated electron-density profile can reach zero. Passing these values into the Sauter calculation results in log(0), producing non-finite values and causing the model evaluation to fail.

This is particularly problematic for because an invalid intermediate trial point can terminate or destabilise the optimisation even when a valid equilibrium exists elsewhere in the search space.

solution

The guard is applied only to the local density array used in the bootstrap-current calculation and does not modify the underlying PROCESS density variable or density profile. Itprevents log(0) and subsequent NaN propagation and allows the optimiser to move through otherwise invalid intermediate trial points.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 50.26%. Comparing base (1fcc5ab) to head (7238d39).

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #4717   +/-   ##
=======================================
  Coverage   50.26%   50.26%           
=======================================
  Files         181      181           
  Lines       30629    30630    +1     
=======================================
+ Hits        15395    15396    +1     
  Misses      15234    15234           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[mkovari]

I am sure this is fine for the moment.

The profile code has a lot of places where density or temperature could become negative or where the gradients could become negative. In the future it would be nice to sort this out properly.