From Research to Product Design via Open Source · From Research to Product Design via Open Source OpenFOAM for Industrial CFD Hrvoje Jasak Wikki Ltd, United Kingdom, Germany and

From Research to Product Design viaOpen Source

OpenFOAM for Industrial CFD

Hrvoje Jasak

Wikki Ltd, United Kingdom, Germany and Brazil

Faculty of Mechanical Engineering and Naval Architecture, Uni Zagreb, Croatia

ESI OpenFOAM Conference USA, Framington Hills, 30 October 2018

From Research to Product Design via Open Source – p. 1

Outline

Objective

• Present my work and vision for OpenFOAM: 1993 to 2018 and towards theCFD 2030

Topics

1. About Me

2. Mission / Vision: OpenFOAM as a revolutionary simulation platform

3. Some examples

• Naval Hydrodynamics CFD

• Turbomachinery CFD and harmonic balance• Block-coupled implicit p-U solver

4. Summary


About Me

Hrvoje Jasak: Education and Professional Experience

• First degree: mechanical engineering, University of Zagreb, Croatia 1992

• PhD, Imperial College London 1993-1996

• Senior development engineer, CD-adapco (Siemens), 1996-2000

• Technical director, Nabla Ltd. 2000-2004: in charge of FOAM development

• Consultant on CFD software, numerics and modelling, ANSYS Fluent 2000-2008

Current Work

• Director, Wikki Ltd: software development, support and consulting 2004-

• Professor, University of Zagreb, Croatia 2007-

• Mercator Fellow, TU Darmstadt, 2016-

• Founder and Committee Member, OpenFOAM Workshop: 14 editions (2006)

• Founder and Chief Tutor, NUMAP-FOAM Summer School: 14 editions (2008)

• Coordinating open source OpenFOAM development to allow public contributions

• Teaching, workshops, lectures and seminars, visiting professorships: TU Delft,Chalmers University, Uni Zaragoza, UFRJ, Seoul National University and others


About Me

Role in OpenFOAM Development

• One of two original developers of OpenFOAM software, starting from 1993

• FVM discretisation, polyhedral mesh handling, linear solvers: Jasak PhD 1996

• Error estimation, adaptive mesh refinement, dynamic mesh, automatic meshmotion, topological changes: (sliding, layering); engine CFD

• Parallelism and HPC support: decomposition/reconstruction, comms

• Mesh generation, conversion, manipulation; pre- and post-processing tools

• Turbulence modelling, LES, free surface flows, solid mechanics, visco-elastic

• Finite Element motion solver, finite area method, ODE solvers

• POD, reduced order modelling

• Geometric parametrisation and automatic optimisation


OpenFOAM and CFD in 2020

My work with OpenFOAM in the Last 25 Years: 1993 – 2018

• The vision at the start of the project was to see how good a code can we write

◦ Is commercial CFD numerics/modelling ahead of academic knowledge: No!

◦ Does a commercial environment lead to high-quality software: No!

• The scope for Open Source is present: all CFD knowledge is in public domain

• Focus the world-wide academic and industrial activity in collaborative work

Where We Are Today?

• ANSYS: World-Wide CFD market is no longer growing (oh, really???)

• ISOPE Conference Sapporo, Jun/2018: 58 CFD papers with OpenFOAM

• OpenFOAM present across all CFD applications, industries and research areas

• Custom applications, multi-physics, new frontiers: this is what CFD is for!


Mission and Vision

Mission

• Write the best CFD code in the world, using modern software engineering andquestioning existing (software, modelling, numerics) paradigms

• (Open source deployment follows as a logical consequence)

• . . . and watch its revolutionising effect in the CFD arena

Vision

• Having witnessed the effect of Open Source CFD in practical simulations

1. Focus the academic and industrial research effort to provide a step-change inuse and simulation capabilities

2. Expand the base of competent researchers, developers and users

3. Provide a framework for industrial-academic collaboration, knowledgetransfer, sharing of work/software and understanding of practical requirements

• Strive for excellence: we can do this better!

Events and Consequences

• By far the most widely used CFD/CCM tool of 2010s

• Significant cost reduction of industrial CFD, with increase in simulation volume

• Opening new areas for CFD simulation and integrated work-flows


From Vision to Reality

Some Recent and Ongoing Work: Numerics and Modelling

• Coupled matrix and linear algebra, block-coupled solution of equation sets

• Conjugate coupled multi-domain solvers: implicit conjugate heat transfer

• Scalable and consistent Overset mesh support

• Immersed boundary support (second generation): Immersed Boundary Surface

• Harmonic balance model for turbomachinery and beyond

• Discontinuous Galerkin discretisation: consistent higher-order methods withpolyhedral support and operator-based implementation


From Vision to Reality

Some Recent and Ongoing Work: Applications

• Turbomachinery CFD: rotor-stator interfaces, incompressible and compressibleturbomachinery solver + validation and verification

• Coupled matrix and linear algebra, block-coupled solution of equation sets

• Conjugate coupled multi-domain solvers: implicit conjugate heat transfer

• Complete naval hydrodynamics CFD capability: free surface, forces andmotions, regular and irregular wave modelling, slamming and green water,propulsor and manoeuvring, free sailing, full-scale ship CFD

• Complex solid mechanics modelling: material and geometrical non-linearity, largedeformation, lubricated contact and self-contact, thermal effects

• Fluid-solid interaction modelling: new library by Tukovic

• Fuel cell and battery modelling: detailed electro-chemistry

• Thermo-acoustics modelling: flame instability (custom numerics)

• Multi-phase compressibility, wave impact and green water modelling

• Solidification, phase change, residual stresses in solidifying materials

• Combustion, catalytic reactions, NOx pollutant modelling

• Complex physics in complex geometry: dynamic mesh


Overview of Research Activity

Research Group Members, CFD Group at University of Zagreb

• CFD Research Group attached to the Chair of Turbomachinery

• 2 professors: Prof. Hrvoje Jasak, Prof. Željko Tukovic

• 1 (+ 1) post doctoral researchers; 6 fully funded PhD students

• Regular external (foreign) visitors working with the group: 3-6 months

• Web: http://www.fsb.hr/cfd; YouTube: 8th Floor CFD@FSB

• Approximately 25 (significant) publications per year

Wikki

• Support, contract development, training and routine calculations with FOAM

• Approximately 150 clients world-wide

• Significant engagement with academic and governmental research organisations

• Offices in 4 countries, with partner coverage for specific markets


Naval Hydro Pack


Naval Hydro Pack

Simulation of Resistance, Loading, Sea-Keeping / Global Performance of Naval Objects

• Objective: Predictive simulation of resistance and wave-induced loads and

global performance of off-shore structures

• Modelling of steady resistance and added wave resistance of ship hulls

• Realistic modelling of regular and irregular sea states, including statistics

• Inclusion of propulsion/mooring/riser forces and hydro-elastic response

• Evaluation of regular, but also slamming and impulsive loads, eg. green water

• . . . with minimal simplifications: Computational Fluid Dynamics (CFD)


Naval Hydro Pack

OpenFOAM in Naval Hydrodynamics CFD

• OpenFOAM is an accepted open-source tool for general CFD applications

• Free surface flow modelling capability is present

• . . . but it is not sufficiently robust and accurate for naval hydrodynamicsapplications

• Significant additional functionality is required

◦ General numerics improvements for free surface flows◦ Wave modelling library

◦ Efficient numerical beach / relaxation zone handling

◦ Improved 6-DOF and hydro-mechanical coupling

Naval Hydro Pack by Wikki

• Collaborative development of best-in-class tools for naval hydrodynamics CFD

• The pack provides full functionality for Naval Hydro CFD

• Tutorials with optimal settings and automated execution (meshing to results)

• Public validation and verification benchmark cases: ready to run

• Professional support and consulting services by Wikki


CFD Numerics Improvements

Necessary Numerical Improvements

• Improved free surface capturing: better VOF, implicitly redistanced Level Set

• Handling free surface discontinuity: ghost fluid method, geometric free surface

advection

• Implicit relaxation zones: far-field blending with prescribed flow solution

• Hydro-mechanical coupling: implicit pressure equation to 6-DOF motion coupling

• Overset mesh / immersed boundary method for appendages and rudder motion

• Rapid steady resistance solver: combining steady and transient formulation

• SWENSE solution decomposition solver for accurate wave propagation

• Regular and irregular wave modelling library

• Propulsor modelling procedure: characterising propulsor performance

• 6-DOF Lie Group solver operating directly on the rotation matrix

• Optimal 6-DOF damping for steady sinkage and trim simulations

• Mooring and riser model interface

• Integrated propulsor modelling

• Linearised free surface solver: rapid single-phase free surface simulations


Steady Resistance

Steady Resistance in Calm Water for a Displacement Hull


Steady Resistance

Steady Resistance in Calm Water: KRISO Container Ship (KCS)

• Computer: Single processor Intel I7 4820K, 3.7 GHz, 4 cores , 16 GB RAM

• A converged and accurate resistance force in 30 min on 1 CPU!

Mesh size Drag [N] Simulation Time Converged Forcefor 200 s Simulation Time [s]

600k 41.93 1153 = 19 min 50700k 41.09 1285 = 21 min 50950k 40.35 1752 = 29 min 501.6M 39.93 2996 = 50 min 502.6M 38.91 14249 = 4.0 hrs 125/754.6M 38.58 27888 = 7.7 hrs 125/75

• Computational and experimental uncertainty in sinkage and trim simulations

0.09 0.12 0.15 0.18 0.21 0.24 0.27 0.3Fr

-2.4

-2

-1.6

-1.2

-0.8

-0.4

0

0.4

σ×

102

σ_EFDσ_CFD

horizontal short bar: ± UD

horizontal long bar: ± USN

0.09 0.12 0.15 0.18 0.21 0.24 0.27 0.3Fr

-0.28

-0.24

-0.2

-0.16

-0.12

-0.08

-0.04

0

0.04τo

τ_EFDτ_CFD

horizontal short bar: ± UD

horizontal long bar: ± USN


Sea-Keeping in Oblique Waves

KCS 2.11–C2 seakeeping case, 45◦



KCS 2.11–C4 seakeeping case, 135◦



Total resistance grid uncertainties for 5 cases:

1. Mean value average uncertainty 10%

2. First order harmonic uncertainties less than 3%

Heave grid uncertainties for 5 cases:

1. Mean value uncertainties range from 2% for head waves case to 27% for thequartering waves case

2. First order harmonic uncertainties are less than 2%

Roll grid uncertainties for 5 cases:

1. Mean value: 3% and 7% for bow and quartering waves, respectively

2. Mean value for beam waves is high: 63%–needs further investigation

3. First order harmonic average grid uncertainty approximately 4%.

Pitch grid uncertainties for 5 cases:

1. First order harmonic uncertainties below 2%


Green Sea Loads

Green Sea Loads: Validation and Verification Study

• Green water impact pressure on deck is compared on ten locations

• Static FPSO model, 9 incident waves observed in total

• Performed complete grid and temporal resolution uncertainty study

• Experimental data from H. H. Lee, H. J. Lim & S. H. Rhee: Experimentalinvestigation of green water on deck for a CFD validation database (2012).


Green Sea Loads

H = 13.5 cm, λ = 2.25 m

1 2 3 4 5 6 7 8 9 10Pressure gauge label

0

100

200

300

400

500

600p

max

, P

a

CFDEFD

Pressure peaks


0

50

100

150

200

250

300

P, P

a s

CFDEFD

Pressure integrals

H = 15.0 cm, λ = 3.0 m


0

100

200

300

400

500

600

700

800

pm

ax, P

a

CFDEFD

Pressure peaks


0

50

100

150

200

250

300

P, P

a s

CFDEFD

Pressure integrals


figures/greenSea/wave3ResultsPeaks.eps

figures/greenSea/wave3ResultsIntegral.eps



Multi-Scale Green Sea Modelling

Multi-Scale Modelling for Practical Green Sea Loads

1. Stochastic scale: linear seakeeping in frequency domain -> establish green seaprobability and choose the worst wave spectrum

2. Perform a 3 hour storm seakeeping simulation to detect green sea events andcalculate ship motion in time

3. For the worst green sea event, run a detailed CFD simulation with only part of theship modelled, including detailed deck structures (deck equipment, pipes etc.); usewave and ship kinematics calculated in step 2 as input


Multi-Scale Green Sea Modelling


Full Scale Free Sailing

Simulating Full-Scale Ship Hull Under Self-Propulsion

• Unstructured grids with: 5M, 7M and 9M cells

• Average y+ ranges from 900 to 1 100 between grids (good range for wall functions)

• Grids generated with cfMesh, appropriate refinement regions

• Propeller is modelled as an actuator disc: propeller-related transient flow featuresare not relevant for calculating achieved speed or force balance

• (Moving rudder not required for this case)

Stern and propeller disc Bow refinement


Full Scale Free Sailing

Predicted Ship Speed within 0.3%

• Achieved ship speed falls within two sea trial measurements

• Grid uncertainty is 0.02 knots for the fine grid

• Simulations time: approximately 2 days on 64 cores

0 100 200 300 400 500 600 700 800Time, s

200

250

300

350

400

450

500

|T| &

|R|,

kN

Hull resistancePropeller thrust

600 650 700 750 800300

310

320

Convergence of resistance and thrust Convergence of achieved speed


Naval Hydrodynamics

Bureau Veritas, France

• Seakeeping, slamming and green water on an elastic ship


Turbomachinery CFD


Turbomachinery CFD

Support for Turbomachinery CFD in foam-extend

• Complex rotor-stator interfaces, for all physics models: GGI, mixing plane

• Incompressible and compressible turbulent flow, MRF and transient

• Complex mesh motion or geometrical scaling (eg pitch-matching)

• Validation and verification: propellers, pumps, turbines, fans, compressors


Harmonic Balance Solver

Harmonic Balance Method

• Harmonic Balance method is a quasi-steady state method developed forsimulation of non-linear temporally periodic flows

• In rotating turbomachinery and other applications, engineering flows exhibit regularperiodicity: wish to describe stable periodical behaviour

• Replacing a transient problem with a set of coupled “steady-state” snapshots byvirtue of using periodicity of the time-signal in:

◦ Boundary conditions

◦ Geometry or relative motion

◦ Flow solution

Harmonic Balance Method: Work-Flow

• Variables are developed into Fourier series in time with n-harmonics andsubstituted into transport equation

• Transport equation with n sine and n cosine parts + mean part is obtained andwritten as a set of 2n+ 1 equations in frequency domain

• Equations are transformed back to time domain in order to be able to usetime-domain boundary conditions and time-domain non-linear flow solver


Harmonic Balance: ERCOFTAC Pump

Harmonic Balance Solver: ERCOFTAC Centrifugal Pump




• Validation of harmonic balance in turbulent incompressible periodic flow

• HB simulations performed using 1 and 2 harmonics: rotor and stator blade count

• Results compared against full transient simulation: excellent agreement

◦ Integral properties: typical error of 2%

◦ Local solution features: pressure on surface in time

◦ Mode and nature of flow instability

• Results are significantly better than expected!

• Substantial reduction in simulation time:◦ Intel Core i5-3570K, 3.4 GHz computer with 16 GB memory◦ Transient run needs approx. 50 blade passages to become quasi-periodic

Transient HB, 1 h HB, 2 h

Simulation time 5 hrs/rotation 52 mins 78 mins

Iterations 600, dt = 5e-5 s 3000 24001 rotation = 0.03 s




-0.2 -0.15 -0.1 -0.05 0x-Axis

-1200

-1000

-800

-600

-400

-200

0

Pre

ssu

re,

Pa

TransientHB, 1hHB, 2h

-0.15 -0.1 -0.05 0 0.05x-Axis

-1200

-1000

-800

-600

-400

-200

0

Pre

ssure

, P

a


-0.15 -0.1 -0.05 0 0.05x-Axis

-1200

-1000

-800

-600

-400

-200

0

Pre

ssure

, P

a


Transient HB, 1h err, % HB, 2h err, % MRF err, %

Efficiency 89.72 88.80 1.0 89.76 0.0 89.65 0.07t = T

3Head 81.48 81.80 0.4 80.45 1.3 84.12 3.14

Torque 0.0297 0.0302 1.7 0.0294 0.9 0.0308 3.57

Efficiency 89.92 88.78 1.3 89.81 0.1t = 2T

3Head 81.48 81.85 0.4 80.6 1.1

Torque 0.0296 0.0302 2.0 0.0295 0.4

Efficiency 89.83 88.85 1.1 89.71 0.1t = T Head 81.49 81.79 0.4 80.39 1.3

Torque 0.0297 0.0302 1.6 0.0294 1.0



Local transient effects comparison, velocity field:

Transient simulation HB with 1 harmonic

HB with 2 harmonics MRF simulation


Water Jet

Water Jet Propulsor: Flow Conditions

• Six-bladed rotor, at 2000 rpm; eight-bladed stator

• Turbulent flow with steady inlet condition, u = 11.43m/s

• No experimental data available: real water jet cavitates at this flow rate

Mesh Layout

• Full annulus with resolved blade tip clearance: 2,153,424 hexahedral cells

• Two domains: rotor and stator connected using a GGI interface

Frozen Rotor MRF Simulation: Coupled Solver

• Rapid and smooth convergence in 150 iterations: 4 hours on a laptop computer

Transient Simulation

• Transient simulation completely impractical due to small mesh size at tip clearancewith large velocities

• Typical ∆t = 1e− 07 s; time for 1 period = 0.03 s

• Transient run ongoing for 4 weeks on a workstation (small mesh)

Harmonic Balance Simulation

• Performing HB simulations with 1, 2 and 7 harmonics


Harmonic Balance: Water Jet

Harmonic Balance for a Water Jet Propulsor


Harmonic Balance: Water Jet

Harmonic Balance for a Water Jet Propulsor

• Temporal variation of head and efficiency: 1 and 2 harmonics

0 0.01 0.02 0.03Time, s

22

22.5

23

23.5

24

Hea

d,

m

HB, 1hHB, 2h

0 0.01 0.02 0.03Time, s

82

83

84

85

86

87

88

89

Eff

icie

ncy

, %

HB, 1hHB, 2h

Water Jet: Future Work

• Further validation & verification work ongoing

• It is possible to extend the HB model to cavitating flow


Coupled Solver

1e-08

1e-07

1e-06

1e-05

1e-04

1e-03

1e-02

1e-01

1e+00

0 500 1000 1500 2000 2500

Res

idual

Iteration

Ux (simpleFoam)

Uy (simpleFoam)

Uz (simpleFoam)

p (simpleFoam)

BiCGStab Ux

BiCGStab Uy

BiCGStab Uz

BiCGStab p

SAMG Ux

SAMG Uy

SAMG Uz

SAMG p


Coupled Solver

Background

• OpenFOAM uses equation mimicking to perform field algebra and discretisation:perfect for simple PDE-s or segregated solution algorithms

• . . . but sometimes we use equation segregation inappropriately

• There exists a family of problems that cannot be solved efficiently withoutinter-equation coupling; some simulations “that work” can be performed 10-s or100-s of times faster than with equivalent segregated algorithms

Objective

• Implement flexible and efficient block-coupled solution infrastructure

• Re-use all operator-based discretisation schemes, parallelisation and boundarycondition tools already available in OpenFOAM

• Optimise top-level code for efficient execution and ease of assembly

Examples

• Incompressible steady pressure-velocity system (with turbulence)

• Compressible multi-phase free surface simulations: under-water explosions

• Complex chemistry: species coupling for high-CFL solution

• Multi-phase and poly dispersed flow; visco-elasto-plastic fluids

• Neutronics: reactor modelling


Coupled Solver

Turbulent Steady Incompressible Flows: SIMPLE or Coupled System

• Equation set contains linear p-U and non-linear U-U coupling

∂u

∂t+∇•(uu)−∇• (ν∇u) = −∇p

∇•u = 0

• Traditionally, this equation set is solved using the segregated SIMPLE algorithm

◦ Low memory peak: solution + single scalar matrix in peak storage

◦ p-U coupling is handled explicitly: loss of convergence (under-relaxation)

◦ Number of iterations is substantial; not only due to non-linearity

◦ Convergence dependent on mesh size: SIMPLE slows down on large meshes

• Block-implicit p-U coupled solution

◦ Coupled solution significantly increases matrix size: 4 blocks instead of 1

◦ . . . but the linear p-U coupling is fully implicit!

◦ Iteration sequence only needed to handle the non-linearity in the U-equation

◦ Net result: significant convergence improvement (steady or transient) at acost of increase in memory usage: reasonable performance compromise!


Coupled Solver

Block Version of Selective Algebraic Multigrid

• Major performance jump: block-coupled AMG with Selective Coarsening

• The algorithm follows the principles of the classical SAMG (Stüben), but uses aprimary matrix (Clees) for coarsening and calculation of interpolation

• Additionally, using additive correction (Hutchinson 1988) for solution acceleration

• Algorithm is extended to non-M-matrices and block coefficients

• New smoother based on Crout’s lower-upper factorisation

• Parallelisation with the in-depth matrix fetch across coupled interfaces

• Support for non-trivial coupling: GGI interface, mixing plane

Block-coupled k − ǫ and k − ω SST turbulence models

• Turbulence equations solved in a single block-coupled system

• Analysis of source terms to establish favourable cross-equation coupling

• Implemented in Diploma Thesis assignment: Robert Keser, Uni Zagreb


Coupled Solver

Performance of the Coupled p-U Solver: Speed and Robustness


Coupled Solver

Performance of the Coupled p-U Solver: Submarine Flight, 14M Cells


Coupled Solver

DrivAer Geometry: External Aerodynamics, Coupled Solver, 13.2M Cells


Coupled Solver

Water Jet: Steady-State Frozen Rotor, MRF Solution, Coupled Solver: ConvergenceHistory


Coupled Solver

Water Jet: Steady-State Frozen Rotor, MRF Solution, Coupled Solver


Summary

Summary

• OpenFOAM and Open Source software infrastructure allows us to leverageexisting technology and deliver custom solution to clients, either via toolsdevelopment or by developing, implementing and validating new physical andnumerical models

• A two-stage approach is needed: academic research + industrial deployment

• (In my work):◦ Wikki handles contracts where industrial collaboration is sought, either

through support, custom development, training, process integration orcollaborative simulation work

◦ Uni Zagreb has the capability of providing research support via industrialcollaboration projects, funded or joint PhD research projects or directcollaboration

• Strength of OpenFOAM is in collaborative Open Source: careful management isneeded to secure the future


NUMAP-FOAM Summer School

NUMAP-FOAM Summer School 2019

• 14th Edition of NUMAP-FOAM Summer School: 19-30/Aug/2019https://www.fsb.unizg.hr/numap

The idea of the Summer School is to expand the physical modellingknowledge, numerics and programming skills of attendees usingOpenFOAM in their research through direct supervision and one-to-onework.

This is NOT an introductory OpenFOAM course: significantunderstanding of the project and software is a pre-requisite forapplication.

• The School accepts 10-15 attendees bringing their own projects to the School overa period of 10 working days

• Work is embedded in the research group with 4–6 tutors providing daily one-to-oneattention

• School is open to “young researchers” (typically PhD students) but also toindustrial users, government labs and professors

• Strong follow-up collaboration and extensive publication lists

• Approx 170 attendees to NUMAP-FOAM, from the start in 2008


Documents

From Research to Product Design via Open Source · From Research to Product Design via Open Source OpenFOAM for Industrial CFD Hrvoje Jasak Wikki Ltd, United Kingdom, Germany and