INFRASTRUCTURE

MINING & METALS

NUCLEAR, SECURITY & ENVIRONMENTAL

OIL, GAS & CHEMICALS

Computer Simulation of Liquid-Solids Slurries for Wastewater Treatment 2016

48 hydrolink number 2/2016

Applied Simulation and Analysis; Bechtel Nuclear, Security, & Environmental; Reston, Virginia 20190

Processing of liquid-solids slurries for waste-water treatment involves handling of dissolved solids and undissolved solids with readily suspended to rapidly settling behaviors. Given a significant loading of dissolved or readily suspended solids, the effective carrier-fluid rheology may exhibit complicated non-Newtonian effects. A simulation-based assessment of wastewater treatment requires a

sophisticated computational fluid dynamics(CFD) flow code with submodels sufficient toaddress this potentially diverse range ofphysics. Reynolds-Averaged Navier-Stokes(RANS) models are the current workhorse.

Simulation is always limited by available compu-tational resources and physics parameteriza-tions. With advances in computationalengineering in parallel processing environmentsand physics submodel development forcomputer simulation codes, many limitations areeither being removed or are being moved to

higher-order details. CFD-RANS models arenow able to meet challenges for simulatingliquid-solids slurry flows in complicated configurations.

Industrial wastewater may contain a significantfraction of undissolved solids with potentiallybroad particle size and density distributions.Granular-Eulerian multiphase modeling is anexample of a CFD-RANS technology that hasbeen formulated to handle this kind of appli-cation. In a Granular-Eulerian multiphase model,each gas, liquid, or solids constituent is treated

COMPUTER SIMULATION OF LIQUID-SOLIDS SLURRIES FORWASTEWATER TREATMENTBY L. JOEL PELTIER, KELLY J KNIGHT, BRIGETTE ROSENDALL, SANJEEB PAL, ANDRI RIZHAKOV, ANDREI SMIRNOV AND CHANTHY IEK

Figure 2. Flow of a Herschel-Bulkley Fluid in a Pipe: Flow (left), Velocities (Right)

Figure 1. Multiphase mixing in an industrial process vessel: Flow (left), Velocities (Right)

49hydrolink number 2/2016

IAHR

as a separate continuous modeling phase.Submodels are used to parameterize interac-tions and behaviors at boundaries.

Comparisons of computational results (from acommercial CFD code CD-adapco/Star-CCM+)to experimental data show the fidelity that canbe achieved. Figure 1 (left) is an instant from asimulation of mixing of a polydisperse loadingof undissolved solids in a Newtonian carrierfluid. The mixing is performed in a vessel priorto the next step of the treatment process. Thetotal solids loading in the vessel is 10% byweight. Approximately half of the solids arereadily suspended. The upper part of the vesselis gas. The particle distribution is characterizedby 6 solids phases with representative particlesizes and densities. The simulation presented inthis article models an existing physical modelexperiment of the mixing of the waste in thevessel. The simulation geometry is derived froma CAD model of the experimental apparatus. Inboth the simulation and the experiment, timehistories of velocity are sampled at six points inthe bulk flow with the velocity samplinglocations at lower, mid, and upper levels. Threelocations provide velocities near the vesselcenterline. Three locations provide velocitiesnear the vessel outer wall. Comparisons of theCFD-RANS predicted velocities to the experi-mental data, Figure 1 (right), confirm modelfidelity to real-world physics.

Dissolved and undissolved readily-suspendedsolids in industrial mixing vessels and otherliquid-solids slurries may be modeled using aneffective fluid rheology and density.Contemporary CFD solvers include a broadrange of rheology submodels, a non-NewtonianHerschel-Bulkley fluid being an example.

In a Herschel-Bulkley fluid, the apparentviscosity of the fluid depends on the local shearrate. In regions of high local shear rates, aHerschel-Bulkley fluid behaves like a Newtonianfluid. As local shear rates reduce, a Herschel-Bulkley fluid becomes more viscous. The localshear rate in a turbulent flow occurs in the dissi-pation range of turbulence. CFD-RANSsolutions provide energy-containing-range(mean-field) statistics, not dissipation rangestatistics. Without an appropriate model linkinglocal shear rates to mean-field statistics, CFDsimulations of Herschel-Bulkley fluids are welldefined only for laminar flows where the dissi-pation range can be resolved explicitly.

IAHR

L. Joel Peltier is a PrincipalEngineer and manager ofComputational Fluid Dynamics (CFD). He specializes in applied

fluid mechanics, CFD, and numerical heattransfer applied to nuclear waste treatmentand industrial processes. He is a member ofASME and Vice-Chair of the ASME Verifica-tion & Validation 20 Code and StandardsCommittee.

Kelly J Knight is a PrincipalEngineer and manager ofAdvanced Simulation andAnalysis (ASA). Kelly spe-cializes in simulation of

fluid and structural mechanics includingmodeling of blast protection, structuralstress/strain and fatigue, multiphase flows,shock and blast wave events and thermalheat transfer. She is a member of the ASME.

Brigette Rosendall is aComputational Fluid Dynamics Specialist forSolar Turbines in SanDiego. She specializes in

solving problems involving multi-phase flow,combustion, radiation, species transport,free-surface, and moving boundaries. She isa member of the AIChE.

Sanjeeb Pal is a Senior Engineering Specialist inComputational Fluid Dynamics. He startedworking for Bechtel

Corporation in 2008. His work in the ASAgroup is focused on Computational Fluid Dynamics application to address engineeringchallenges on Bechtel EPC Projects. He is amember of the ASME.

Andri Rizhakov is a SeniorEngineer in ComputationalFluid Dynamics. Hespecializes in nuclear wastetreatment, multiphase flows

for industrial applications, and analytics workvia machine learning paradigms. Andri’sleadership activities include team guidanceinvolving writing of CFD quality processprocedures. He is a member of the ASME.

Andrei Smirnov is a SeniorEngineering Specialist inComputational FluidDynamics. He specializes indesign of nuclear reactors,

nuclear waste treatment, multi-phase flowanalyses for industrial applications, andcombustion software development. He is amember of the ASME and the ANS and a holder of one US patent.

Chanthy Iek is anEngineering Specialist inComputational FluidDynamics. He specializes in

shock wave dynamics and radiation/boiling/particulate deposition for chemical processes.His engineering experiences include CFD,turbomachinery, aero-acoustics, and air-breathing and rocket propulsion. He is amember of the ASME.

A method to extend CFD modeling of Herschel-Bulkley fluids into the turbulence regime wasrecently presented at the Star-CCM+ GlobalConference (Peltier et al, 2016). This modelextension uses turbulence theory to estimaterepresentative local maximum shear ratemagnitudes from CFD-RANS data enablingsimulations of Herschel-Bulkley fluids in theturbulence regime.

Figure 2 (left) shows CFD predicted viscositiesfor flow in a pipe of a Herschel-Bulkley fluid inthe laminar, transitional, and turbulenceregimes. The slice shown is from the pipecenterline to the upper outer wall. Comparisons

of the CFD predicted velocities to experimentaldata, Figure 2 (right), confirm model fidelity toreal-world physics.

The examples shown for simulation of liquid-solids slurries underscore that capabilities ofcontemporary commercial CFD flow codes arerapidly advancing and support a conclusionthat a simulation-based assessment of waste-water treatment is possible with an expectationfor fidelity to real-world physics. n

ReferencesPeltier, J., Rizhakov, A., Rosendall, B. Inkson, N., and Lo, S., 2015:

Evaluation of RANS Modeling of Non-Newtonian Bingham Fluidsin the Turbulence Regime using STAR-CCM+®, STAR-CCM+Global Conference Location: San Diego, CA, March 15-18, 2015.