rtx-cfd: PatchConvection::TvdVanAlbada — van Albada deferred correction on the curvilinear predictor (downwind-side linear weight, gradient-ratio r over the face d lengths, far-upwind across the opposite face, boundary faces upwind); annulus MMS orders 2.10/1.69 at 0.24× upwind; cylinder-flag MMS orders 1.98/1.97 (1.06× upwind — diffusion-dominated, recorded); knobs RTX_OVERSET_CFD1_TVD, RTX_OVERSET_MAX_ROUNDS, RTX_CF_SCHEME

Co-Authored-By: Claude Fable 5.1 <[email protected]>
Claude-Session: https://claude.ai/code/session_01X2GmJXeQ2njUecEKiJZ1G2

crates/specialized/rtx-cfd/src/solvers/incompressible/curvilinear/mod.rs

@@ -88,6 +88,14 @@ pub enum PatchConvection {
     Upwind,
     /// No convection: the Stokes limit, for the second-order MMS gate.
     None,
+    /// Deferred-correction TVD with the van Albada limiter (the harness's
+    /// scheme): the upwind face value plus `w_up psi(r) (phi_dn − phi_up)`,
+    /// `w_up` the mesh's linear weight of the downwind side and `r` the
+    /// ratio of the two one-sided gradients (so a linear field on a
+    /// stretched row gives `r = 1` and the mesh's own linear face value).
+    /// Faces whose far-upwind cell lies outside the patch fall back to
+    /// upwind. Explicit, like the rest of the predictor's convection.
+    TvdVanAlbada,
 }
 
 /// How the across-patch diffusion is time-stepped.