Flow Simulation
FlowDict
The FlowDict module simulates flow experiments and post-processes the simulation results to predict flow velocity, flow permeability, and flow resistivity as effective material properties.
A flow experiment in FlowDict requires as input:
- 3D representation of a structure or material
- Newtonian fluid (gas or liquid) with constant density (incompressible)
- Experimental process parameters, such as mass flow rate, pressure difference and flow direction
Examples of applications
- Determine air and water permeability in woven fabrics
- Study gas and liquid permeability, as well as pressure drop, in filter media
- Predict volumetric flow rates and pressure in complete filters
- Predict the permeability for gas extraction in gas reservoirs
- Characterize flow properties of groundwater in aquifers
- Predict absolute permeability of digital rocks for tertiary oil recovery
Flow Simulation in Fiber Structures
- Prediction of pressure drop and complete
permeability tensor - Determination of layer-wise pressure and velocity
in the pore space along the direction of flow (Fig. a)
(a) Layer-wise pressure along the direction of flow
(b) A model of a fiber structure designed with FiberGeo
(c) Water velocity in a fiber structure - simulated using FlowDict
and the LIR solver in FlowDict
(d) Black boxes represent the adaptive mesh grid of the LIR solver
Flow Simulation in Filter Pleats
- Prediction of pressure drop in oil filter pleats
- Identification of critical areas with high velocities
- Allows optimization of the shape and quantity of pleats
(a) Oil filter pleats created using PleatGeo
(b) Flow velocity through oil filter pleats (2D)
(c) Flow velocity through oil filter pleats (3D)
Flow Simulation in Digital Rocks
- Prediction of absolute permeability
- Support for porous voxels enables multiscale simulation
(a) Model of a Berea sandstone
(b) Flow through a Berea sandstone
Flow Simulation in Complete Filters
- Prediction of pressure and flow rates in complete filters
- Allows multiple inlet and outlet zones
- Identification of critical areas with high velocities
(a) Flow velocity through a complete oil filter, including inlet and outlet
(b) Pressure distribution in a complete oil filter
FlowDict performs five categories of calculations for the prediction of:
- Mean flow velocity for a given pressure drop
- Pressure drop for a given mean flow velocity
- Full or partial permeability tensor
- Volumetric flow rates for given pressures at the inlet and outlet
- Pressure for given volumetric flow rates at the inlet and outlet
In the post-processing, FlowDict calculates the permeability of the material according to Darcy's principle from the mean flow velocity, liquid viscosity, pressure drop, and media thickness.
Darcy's principle is only applicable to very slow flows (creeping or Stokes flows) with a Reynolds number close to zero.
Faster flows, in which the relationship between pressure drop and mean velocity is non-linear, are described by the Navier-Stokes equations.
For slow and faster flows, FlowDict assumes a steady flow regime without time-dependent behavior, such as turbulence.
In such a steady flow regime, neither mean flow velocity nor pressure drop can become arbitrarily large. But, for very fast flows where no steady-state solution exists, the pressure drop and mean flow velocity can be predicted using the Forchheimer approximation in FlowDict.
FlowDict allows the calculation of steady-state flows for the following equations (see Solver Technologies). The following refers to pore voxels filled with a fluid, solid voxels representing obstacles to the flow, and porous voxels representing an underlying microstructure through their permeability value.
- Stokes (with EJ, SimpleFFT and LIR solvers):
- Applies to slow flows with a Reynolds number close to zero
- The geometry must be fully resolved in pore and solid voxels
- Stokes-Brinkman (with SimpleFFT and LIR solvers):
- Applies to slow flows with a Reynolds number close to zero.
- The geometry consists of pore, solid, and porous voxels.
- Navier-Stokes (with SimpleFFT and LIR solvers):
- Applies even to fast flows with high Reynolds numbers
- The geometry must be fully resolved in pore and solid voxels
- Navier-Stokes-Brinkman (with SimpleFFT and LIR solvers)
- Applies even to fast flows with high Reynolds numbers.
- The geometry consists of pore, solid, and porous voxels
- Darcy (with EJ and LIR solvers)
- Applies to slow flows with a Reynolds number close to zero
- The geometry mainly consists of porous and solid voxels
- Ideal for digital rocks obtained through medical-grade CT scans
These modules are often combined with FlowDict:
| Image processing & Image analysis | ImportGeo-Vol | |||||
| Characterization & Analysis | GrainFind | FiberFind | PoroDict + MatDict | |||
| Modelling & Design | GrainGeo | FiberGeo | FoamGeo | PleatGeo | WeaveGeo | PaperGeo |
| Simulation & Prediction | FilterDict | AddiDict | AcoustoDict | SatuDict |
Which modules suit you best depends on the nature of your application.