Splitter Erosion: DPM

Splitter Erosion CFD Simulation Training using DPM by ANSYS Fluent

Description

A splitter is a device used to uniformly distribute incoming fluid flow through outlets of matching shape and size. Beyond evenly dividing flow rate, splitters can incorporate filtration to remove impurities and improve outlet gas purity. Impurities such as sand and various metal oxides can cause progressive erosion on equipment surfaces over time, making erosion analysis on transmission pipelines and flow-distribution equipment a critical engineering concern.

Using the Discrete Phase Model (DPM), this study examines how impurities within a working fluid affect erosion on a gas splitter body. The impurity-laden gas enters vertically at 5 m/s and is directed through three outlet nozzles. Impurity distribution, concentration, adsorption, and reflection behavior within the installed filters were analyzed using ANSYS Fluent, applying multiple erosion models to accurately predict erosion effects under varying operating conditions.

The splitter geometry features three outlet nozzles, with a mainstream inlet diameter of 1.6 cm and outlet nozzle diameters of 0.3 cm. Filter fins measuring 2.5 cm in length are positioned inside the geometry, built using Design Modeler. The domain was meshed in ANSYS Meshing using an unstructured grid of 2,728,426 elements, with curvature and proximity refinement applied near the fins, and boundary layer meshing along the walls to satisfy turbulence model Y+ requirements.

Methodology

The governing equations were solved using ANSYS Fluent's pressure-based, steady-state solver, with gravitational effects excluded. Discrete phase particles were tracked using RANS with a Lagrangian reference frame, where particle inertia is balanced against the forces acting on each particle. Given the high-speed internal flow within the domain, natural gas density was treated as constant, with relevant thermodynamic properties — viscosity, thermal conductivity, and impurity density — defined accordingly.

Key simulation parameters included:

• Natural gas properties: density of 0.65 kg/m³, viscosity of 0.00013 kg/m·s

• Impurity particles: density of 1600 kg/m³, uniform diameter of 0.15 mm, total flow rate of 0.04627 kg/s

• DPM settings: 10 continuous-phase iterations per DPM step, maximum step tracking of 50,000, trapezoidal tracking scheme, spherical drag law, and stochastic turbulent dispersion via the Discrete Random Walk model

• Particle tracking outcome: of 24,900 tracked particles, 8,202 were trapped and 16,695 escaped

• Boundary conditions: 5 m/s velocity inlet, 0 Pa gauge pressure outlet, trap condition on fin walls, escape condition on external domain walls

• Turbulence model: Realizable k-ε with enhanced wall treatment

• Solution methods: SIMPLE pressure-velocity coupling, standard pressure discretization, second-order upwind for momentum, and first-order upwind for turbulent kinetic energy and dissipation rate

Four erosion models were evaluated: the Generic model (broadly applicable, given sand is a common impurity across most cases), the Finnie model (empirically based, suited to malleable materials and sensitive to collision angle and velocity), the Oka model (accounts for wall hardness, making it well-suited to transmission pipe erosion analysis), and the McLaury model (intended for suspended solids in water, and found unsuitable for this particular case).

Analysis

Erosion contours across all applicable models consistently showed that particle impact on the upper wall — driven by high fluid velocity — produces greater erosion than other regions, with the outlet nozzle walls also experiencing elevated erosion. Oka erosion diagrams were monitored throughout the solution process to help assess convergence behavior.

Impurity concentration contours revealed that near the splitter's outlet, high downstream velocity combined with a reduced cross-sectional area made particle exit difficult, leading to particle accumulation and increased impurity concentration in that region. However, the filter fins were shown to enhance impurity particle adsorption, as reflected in the trap-versus-escape particle tracking results.