NPSS Convergence Limits for Stalled or Oscillating Residuals

Modeling a steady-state single-spool turbojet in NPSS requires the modified Newton-Raphson solver to converge reliably, yet the choice of remedy remains unclear when iteration stalls or oscillates. The documented controls include residual monitors and solver-side parameters, but no deterministic procedure dictates whether to adjust initial guesses, reorder elements, or tune relaxation first.

Because convergence depends heavily on deck structure and starting values, modelers need to know whether NPSS defines behavioral limits or diagnostic thresholds that prioritize one approach over another. Without such guidance, each non-convergent deck becomes a manual search rather than a documented workflow.

Does NPSS documentation specify criteria for choosing relaxation tuning over element reordering? Are there iteration limits or residual patterns that indicate a deck will not converge without structural changes? How do documented solver behaviors differ between steady-state and transient modes when managing stalled residuals?