conatus.expert.eigenvalues

conatus.expert.eigenvalues(plan, group=None, *, reduced=None, method=None, target=None, n_modes=None, left=False, drift_plan=None, drift_tol=REQUIRED, spurious='tail', region=None)

Eigenvalues sigma of (L + sigma M) x = 0 in one group of plan.

The rank report runs first (refuses ill-posed systems). method is ‘dense’ (all modes, QZ) or ‘shift-invert’ (n_modes modes nearest target, default DEFAULT_N_MODES; k is its deprecated name: k is a wavenumber everywhere else). With drift_plan (a finer-resolution plan of the same problem) only modes that reappear there within drift_tol are kept, and the count of removed modes is logged in the result. drift_tol has no default: required with drift_plan.

spurious: "tail" (default) flags the modes whose eigenvector has its energy in the last wall-normal coefficients (spurious tau modes; declared tolerances conatus.tolerances.SPURIOUS_TAIL_TOL): they are kept, listed by explain() and ranked last by order(); "none" switches the flag off (said by the result).

region: a cn.box(re=(a, b), im=(c, d)) of the eigenvalue plane: every eigenvalue of the group in it, counted by the argument principle and computed by Beyn’s contour-integral method, each certified by its residual (conatus.contour.RegionEVPResult: count, count_method, multiplicity); method is then "contour". method=None (the default) means "dense" without a region and "contour" with one; any other method given with region= ("dense" too: it would have been replaced without a word) is refused, as are target and left=True.

Parameters:

• plan (Plan)

• reduced (bool | None)

• method (str | None)

• target (complex | None)

• n_modes (int | None)

• left (bool)

• drift_plan (Plan | Callable[[], Plan] | None)

• drift_tol (float)

• spurious (str)

Return type:

EVPResult