Skip to content
SEINA-3D

What the module models

During premixing and steam explosion, molten core material coexists in distinct topological forms alongside liquid water and steam: a continuous melt jet entering from the release nozzle, coarse melt droplets stripped from the jet, and sub-millimetre fragments produced by rapid fragmentation. In SEINA-3D v1, the five-field Eulerian formulation represents liquid coolant (l\mathrm{l}), vapor (v\mathrm{v}), the connected melt jet (j\mathrm{j}), coarse melt droplets (d\mathrm{d}), and fine fragments (f\mathrm{f}) on a shared three-dimensional staggered Cartesian grid. From initial pool conditions and nozzle inflow parameters, the module evolves phase volume fractions, velocities, temperatures, and sharp gas–liquid or jet boundaries across the vessel.

Governing equations

Within every control volume, the five volume fractions αk\alpha_k (k∈{l,v,j,d,f}k \in \{\mathrm{l}, \mathrm{v}, \mathrm{j}, \mathrm{d}, \mathrm{f}\}) satisfy the volume constraint:

∑k∈{l,v,j,d,f}αk=1\sum_{k \in \{\mathrm{l}, \mathrm{v}, \mathrm{j}, \mathrm{d}, \mathrm{f}\}} \alpha_k = 1

Mass conservation for each field kk accounts for convective transport and paired interfield mass transfer Γmk\Gamma_{mk} (Γmk=−Γkm\Gamma_{mk} = -\Gamma_{km}), including evaporation and condensation between water and steam, jet stripping into droplets, and droplet fragmentation into fine fragments:

∂(αkρk)∂t+∇⋅(αkρkuk)=∑m≠kΓmk\frac{\partial (\alpha_k \rho_k)}{\partial t} + \nabla \cdot (\alpha_k \rho_k \mathbf{u}_k) = \sum_{m \neq k} \Gamma_{mk}

All fields share a common pressure pp within each cell. Momentum conservation for each velocity field rr includes pressure gradient, viscous stress τr\boldsymbol{\tau}_r, gravity, interfield drag Msr\mathbf{M}_{sr}, momentum carried by mass transfer, and capillary force Fσ,r\mathbf{F}_{\sigma,r}:

∂(αrρrur)∂t+∇⋅(αrρrurur)=−αr∇p+∇⋅(αrτr)+αrρrg+∑s≠rMsr+∑s≠rΓsrusr+Fσ,r\begin{aligned} &\frac{\partial (\alpha_r \rho_r \mathbf{u}_r)}{\partial t} + \nabla \cdot (\alpha_r \rho_r \mathbf{u}_r \mathbf{u}_r) \\ &\quad = -\alpha_r \nabla p + \nabla \cdot (\alpha_r \boldsymbol{\tau}_r) + \alpha_r \rho_r \mathbf{g} \\ &\qquad + \sum_{s \neq r} \mathbf{M}_{sr} + \sum_{s \neq r} \Gamma_{sr}\mathbf{u}_{sr} + \mathbf{F}_{\sigma,r} \end{aligned}

Specific internal energy eke_k evolves with reversible work from volume-fraction changes and velocity divergence, enthalpy HmkH_{mk} carried by phase or morphology transfer, and interfacial heat transfer QmkQ_{mk}:

∂(αkρkek)∂t+∇⋅(αkρkekuk)=−p[∂αk∂t+∇⋅(αkuk)]+∑m≠kΓmkHmk+∑m≠kQmk\begin{aligned} &\frac{\partial (\alpha_k \rho_k e_k)}{\partial t} + \nabla \cdot (\alpha_k \rho_k e_k \mathbf{u}_k) \\ &\quad = -p\left[\frac{\partial \alpha_k}{\partial t} + \nabla \cdot (\alpha_k \mathbf{u}_k)\right] \\ &\qquad + \sum_{m \neq k} \Gamma_{mk} H_{mk} + \sum_{m \neq k} Q_{mk} \end{aligned}

For dispersed melt fields q∈{d,f}q \in \{\mathrm{d}, \mathrm{f}\} interacting with continuous coolant fields c∈{l,v}c \in \{\mathrm{l}, \mathrm{v}\}, interfield drag is expressed as Mqc=Cqc∣uq−uc∣(uq−uc)\mathbf{M}_{qc} = C_{qc}|\mathbf{u}_q - \mathbf{u}_c|(\mathbf{u}_q - \mathbf{u}_c) with exchange coefficient:

Cqc=3αq4DqCdρcC_{qc} = \frac{3 \alpha_q}{4 D_q} C_{\mathrm{d}} \rho_c

where DqD_q is the Sauter mean diameter of field qq and CdC_{\mathrm{d}} follows the Ishii–Mishima flow-regime drag constitutive relations across bubbly, transition, and droplet-mist regimes. Interfacial closures for stripping, boiling, and phase change are detailed in Jet breakup, droplet fragmentation, and boiling (SEINA-3D v1).

Numerical method

Large-scale interfaces of the liquid pool and connected melt jet are tracked using the Volume of Fluid method with Piecewise Linear Interface Calculation (VOF–PLIC):

∂αk∂t+∇⋅(αkuk)=1ρk∑m≠kΓmk−αkρkDρkDt\begin{aligned} \frac{\partial \alpha_k}{\partial t} + \nabla \cdot (\alpha_k \mathbf{u}_k) &= \frac{1}{\rho_k}\sum_{m \neq k} \Gamma_{mk} - \frac{\alpha_k}{\rho_k}\frac{\mathrm{D}\rho_k}{\mathrm{D}t} \end{aligned}

Within each interface cell, a planar segment n⋅x+d=0\mathbf{n}\cdot\mathbf{x} + d = 0 is reconstructed from local volume-fraction gradients to match αk\alpha_k exactly before directional geometric flux advection. Surface tension on reconstructed interfaces is evaluated with the Continuum Surface Force (CSF) formulation:

Fσ=σκ∇α,κ=−∇⋅(∇α∣∇α∣)\mathbf{F}_{\sigma} = \sigma \kappa \nabla \alpha, \quad \kappa = -\nabla \cdot \left(\frac{\nabla \alpha}{|\nabla \alpha|}\right)

Scalar quantities (p,αk,ρk,ekp, \alpha_k, \rho_k, e_k) reside at cell centres while normal velocity components sit on staggered cell faces, coupled at each time step through the multiphase pressure–velocity–energy solver.

Verification

Three classical two-phase flow benchmarks verified interface advection, free-surface impact, and capillary dynamics in Sun et al. (2027):

  • Ransom one-dimensional faucet flow: tested on three axial grids (9696, 192192, and 384384 cells) at Courant number CFL=0.38\mathrm{CFL} = 0.38. At t=0.6 st = 0.6\ \mathrm{s}, the finest grid captured the analytical void-wave front (7.7658 m7.7658\ \mathrm{m}) within 0.070%0.070\% relative error and matched the steady outlet liquid mass flow rate within 0.16%0.16\%.
  • SPHERIC Test 2 three-dimensional dam break: simulated on a 216×68×68216 \times 68 \times 68 fine grid against Kleefsman’s obstacle-impact experiments, yielding a median normalized root-mean-square error (NRMSE) of 0.1590.159 across wave gauges H1–H4, a median NRMSE of 0.1390.139 across wall pressure sensors P1–P8, and a first-impact peak of 10.11 kPa10.11\ \mathrm{kPa} at sensor P1 compared with the measured 11.27 kPa11.27\ \mathrm{kPa} (−10.2%-10.2\%).
  • Three-dimensional rising bubble: computed on a 72×72×14472 \times 72 \times 144 grid at Eötvös number Eo=9\mathit{Eo} = 9, preserving total gas volume within 0.01%0.01\% over 3.0 s3.0\ \mathrm{s} while matching reference bubble centroid height and terminal rise velocity within 0.47%0.47\% and 0.76%0.76\%.