Modelling Fiber Suspensions Flows Using Rheological Data

38
Modelling Fiber Suspensions Flows Using Rheological Data M. Graça Rasteiro COSt Action FP1005 meeting Nancy, 13-14 October 2011 Portugal

description

Modelling Fiber Suspensions Flows Using Rheological Data. COSt Action FP1005 meeting Nancy, 13-14 October 2011. M. Graça Rasteiro. Portugal. Overview. UNIVERSITY of COIMBRA. Objectives Overview of work being developed Rheological characterization Pilot rig CFD model Results - PowerPoint PPT Presentation

Transcript of Modelling Fiber Suspensions Flows Using Rheological Data

Page 1: Modelling Fiber Suspensions Flows Using Rheological Data

Modelling Fiber Suspensions Flows Using Rheological Data

M. Graça Rasteiro COSt Action FP1005 meeting Nancy, 13-14 October 2011

Portugal

Page 2: Modelling Fiber Suspensions Flows Using Rheological Data

Objectives

Overview of work being developed

Rheological characterization

Pilot rig

CFD model

Results

Future work

OverviewUNIVERSITY of COIMBRA

Page 3: Modelling Fiber Suspensions Flows Using Rheological Data

Objectives

Flow properties of pulp suspensions are important for the optimization of most unit operations in pulp and paper making. Therefore, it is necessary to understand the specific hydrodynamic features of fibre suspensions.

UNIVERSITY of COIMBRA

Page 4: Modelling Fiber Suspensions Flows Using Rheological Data

Plug Flow

Mixed Flow

Turbulent Flow

Pulp suspensions flowing in pipes exhibit three basic types of shear flow mechanisms:

Objectives

water

Plug flow

Turbulent

flow

Transition

flowL

P

maxV WTV WV redV

UNIVERSITY of COIMBRA

Page 5: Modelling Fiber Suspensions Flows Using Rheological Data

Objectives

Construction of a flow model able to predict the flow behaviour of pulp fibre suspensions represents an important step in this area.

The Turbulence Model is one of the simplest and most used turbulence models for industrial applications. Modifications to the

k- ε model can lead to an efficient description of the flow of concentrated fiber suspensions.

k

UNIVERSITY of COIMBRA

Strategy: Pseudo-Homogeneous Model

Knowledge of the rheological behaviour is essential for the construction of a realistic model.

Page 6: Modelling Fiber Suspensions Flows Using Rheological Data

Overview of the research developed

Rheological characterization of fiber suspensions (different fiber types and consistencies).

-new rotational rheometer (Searl effect) developed at UCM

UNIVERSITY of COIMBRA

Development of models for the rheology of fiber suspensions.(Carla A.F. Ventura, A. Blanco, C. Negro, F.A.P. Garcia, P. Ferreira, M.G. Rasteiro, "Modelling Pulp

Fibre Suspension Rheology", Tappi J, 6, 7, 17-23 (2007).)

Identification of the main parameters with a stronger impact on the rheology of fiber suspensions.(Carla A.F. Ventura, A. Blanco, C. Negro, F.A.P. Garcia, P. Ferreira, M.G. Rasteiro, "Modelling Pulp

Fibre Suspension Rheology", Tappi J, 6, 7, 17-23 (2007).)

Page 7: Modelling Fiber Suspensions Flows Using Rheological Data

Overview of the research developed

Pilot rig for the study of the flow of fiber suspensions in pipes.

-Flow tests varying fiber type, suspension consistency, temperature, pipe diameter and material.

UNIVERSITY of COIMBRA

Identification of the main parameters with a stronger impact on the flow regimen.(Carla Ventura, Fernando Garcia, Paulo Ferreira, Maria Rasteiro; "Flow Dynamics of Pulp Fibre Suspensions", Tappi J, 7, 8, 20-26 (2008).)

New set of design correlations for friction factor vs Reynolds number based on a statistical design, showing the dependence on consistency and pipe diameter for two flow regimes.(Carla A.F. Ventura, Fernando Garcia, Paulo Ferreira, Maria Rasteiro, "Modelling Pipe Friction Loss of Pulp Fibre Suspensions", CHERD (2011) – in revision.)

Modelling of the flow of fiber suspensions in pipes- turbulent regime – pseudo-homegeneous approach. (Carla A.F. Ventura, Fernando Garcia, Paulo Ferreira, Maria Rasteiro, "Modelling the Turbulent Flow of Pulp Suspensions”, I&ECR, 50,16, 9735–9742 (2011).)

Page 8: Modelling Fiber Suspensions Flows Using Rheological Data

Rheological characterization

New plate rotational Rheometer – Searl effect

UNIVERSITY of COIMBRA

Induces uniform fibre distribution

Measures shear in the rotor (mobile plate) and in the vessel (fixed plate)

Calculates the difference between torque applied by the rotor and torque transmitted by

the fluid to the vessel

1

2

34

5

6

1-Analytical scales;2-Arms to measure torque;3-Rotor;4-Vessel;5-Computer connected to the scales;6-device to control velocity.

5

5

1

6

Page 9: Modelling Fiber Suspensions Flows Using Rheological Data

Rheological characterization

Suspensions tested:

UNIVERSITY of COIMBRA

Pulp TypeFibre length

(mm)

Consistencies %

(w/w)

Recycled pulp 1.14±0.04 1.4 – 4.23

Eucalypt bleachedkraft pulp

0.71±0.03 1.45 – 3.5

Eucalypt (90%) + pine (10%) bleached pulp

0.61±0.06 0.9 – 3.2

Pine unbleached kraft pulp 2.56 ±0.14 0.8 – 3.6

Page 10: Modelling Fiber Suspensions Flows Using Rheological Data

Rheological characterization

Typical Rheograms

UNIVERSITY of COIMBRA

(Nm-2)

C=3.2%C=2.5 %C=2.2 %C=1.9 %

C=1.6%C=1.2 %C=0.9 %

pine + eucalyptus suspension

ap (Pa.s)

Page 11: Modelling Fiber Suspensions Flows Using Rheological Data

Rheological characterization

y - yield stress

k – consistency coefficient

n – flow index

UNIVERSITY of COIMBRA

-Using an experimental design the influence of fibre characteristics (length), consistency and temperature on y were evaluated.

- n an k are mainly influenced by consistency

Herschel-Bulkley model

- Yield stress increases with consistency and fibre length- Temperature has got a negative effect on yield stress

See: Ventura C, Blanco A, Negro C, Ferreira P, Garcia F, Rasteiro M, Tappi J, 6 (7) 17 (2007)

Page 12: Modelling Fiber Suspensions Flows Using Rheological Data

Objective:

Flow model

To model the turbulent flow of pulp fibre suspensions in pipes using CFD(FEM).

UNIVERSITY of COIMBRA

Pseudo-Homogeneous Model

COMSOL Multiphysics Software, version 3.5

Modified k-ε model.

Page 13: Modelling Fiber Suspensions Flows Using Rheological Data

Pilot rig

Pilot Rig

UNIVERSITY of COIMBRA

Page 14: Modelling Fiber Suspensions Flows Using Rheological Data

Pilot rig

Pilot Rig

UNIVERSITY of COIMBRA

Page 15: Modelling Fiber Suspensions Flows Using Rheological Data

Pulp type effect Pipe -3”SS

Experimental Results- pressure drop

0

50

100

150

200

250

300

350

400

0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0

Velocity(m/s)

∆P

(mm

H2O

/m p

ipe)

recycled pulp suspension C=4.2% eucalypt pulp suspension C=3.5 %pine Pulp suspension C=3.4%pulp

UNIVERSITY of COIMBRA

Page 16: Modelling Fiber Suspensions Flows Using Rheological Data

Details of Governing Equations

Equations for the turbulence intensity and length scales

Turbulence damping assumed – I and l were adjusted depending on fiber type

and consistency.

UNIVERSITY of COIMBRA

Equation for ε

2TIk

Equation for k

TL

k 23

81

Re

hDT II

hT DlL

I – turbulence intensity scaling parameter

l - turbulence length scaling parameter

Viscosity was supplied as a function of local shear rate in the pipe cross-section (data extracted from the rheograms).

Page 17: Modelling Fiber Suspensions Flows Using Rheological Data

Results - Modelling

Experimental data

UNIVERSITY of COIMBRA

Pulp Type

Consistencies %

SS Pipes and PE Pipes

3” and 4”

Recycled pulp 1.4 – 4.23

Eucalypt bleached kraft pulp 1.45 – 3.5

Eucalypt (90%) + pine (10%) bleached pulp

0.9 – 3.2

Pine unbleached kraft pulp 0.8 – 3.6

Page 18: Modelling Fiber Suspensions Flows Using Rheological Data

Turbulence parameters values

Pulp type Turbulence Parameters very low consistencies low consistencies

Recycled

Consistency (%) 0.72 0.60 0.61 1.40 1.80 2.30 2.70

I 0.01 0.01 0.01 0.009 0.008 0.007 0.005

l 0.005 0.005 0.005 0.005 0.005 0.005 0.005

Eucalypt

Consistency (%) 0.77 0.91 1.4 1.5

I 0.09 0.09 0.007 0.003

l 0.005 0.005 0.005 0.005

Eucalypt + pine

Consistency (%) 0.71 0.77 0.9 1.2 1.3

I 0.07 0.05 0.01 0.005 0.003

l 0.005 0.005 0.005 0.005 0.005

Pine

Consistency (%) 0.66 0.76 0.8 1

I 0.01 0.01 0.0005 0.0003

l 0.005 0.005 0.005 0.005

UNIVERSITY of COIMBRA Results - Modelling

Page 19: Modelling Fiber Suspensions Flows Using Rheological Data

Results - Modelling

Comparison between predicted and experimental pressure drop (Pa/m) for the turbulent regime

UNIVERSITY of COIMBRA

0

300

600

900

1200

1500

1800

0 300 600 900 1200 1500 1800

Experimental

Pre

dict

ed

Recycled 0.61%

Recycled 0.76%

Recycled 1.4%

Recycled 1.8%

Recycled 2.3%

Recycled 2.7%

0

400

800

1200

1600

2000

0 400 800 1200 1600 2000

Experimental

Pre

dic

ted

Eucalypt 0.77%

Eucalypt 0.91%

Eucalypt 1.4%

Eucalypt 1.5%

0

300

600

900

1200

1500

1800

0 300 600 900 1200 1500 1800

Experimental

Pre

dic

ted

eucalypt+pine 0.71%

eucalypt+pine 0.90%

eucalypt+pine 0.77%

eucalypt+pine 1.20%

eucalypt+pine 1.30%

0

300

600

900

1200

1500

1800

0 300 600 900 1200 1500 1800Experimental

Pre

dic

ted

pine 0.66%

pine o.76%

pine 0.8%

pine 1%

c) eucalypt+pine d) pine

a) recycled b) eucalypt

Page 20: Modelling Fiber Suspensions Flows Using Rheological Data

Results - Modelling

Turbulence intensity scaling parameter versus suspension consistency

UNIVERSITY of COIMBRA

0.0001

0.001

0.01

0.1

1

0.00 0.50 1.00 1.50 2.00 2.50 3.00

Consistency (%)

I

Recycled Pulp

Eucalypt Pulp

Eucalypt/Pine Pulp

Pine Pulp

Low Consistencies

High Consistencies

Page 21: Modelling Fiber Suspensions Flows Using Rheological Data

Future work

▪ Introduce in the model experimental information on the turbulence damping.

▪ Modify the transport equations for k and ε to include mechanistic damping terms.

Establish quantitative correlations for the turbulence intensity and length scales, as a function of fiber characteristics and consistency.

▪ Model the intermediate regime (plug of fibers + water annulus).

Acquire experimental information on the fiber plug evolution with Reynolds number (different fiber types and consistencies), using tomographic techniques.

UNIVERSITY of COIMBRA

Page 22: Modelling Fiber Suspensions Flows Using Rheological Data

Acknowledgments:

RAIZ – Instituto Investigação da Floresta e Papel;

Gopaca, S.A.;

Prado Karton, S.A.;

Soporcel - Grupo Portucel Soporcel;

CELTEJO - Empresa de Celulose do Tejo, S.A.;

European Project NODESZELOSS;

FCT Project FIBERFLOW;

UCM;

Carla Ventura

Carla Cotas

Joy Iglesias

F. Garcia

Paulo Ferreira

UNIVERSITY of COIMBRA

Page 23: Modelling Fiber Suspensions Flows Using Rheological Data

Thank you for your attention

UNIVERSITY of COIMBRA

Page 24: Modelling Fiber Suspensions Flows Using Rheological Data

Conclusions

▪ The pressure drop profiles obtained using COMSOL Multiphysics Software agree very well with the experimental results obtained.

▪ The use of the k-ε Turbulence Model, associated with the rheological data acquired in a specially built viscometer, revealed to be a good strategy for the prediction of pressure drop values for fibre suspension flow.

▪ For very low consistencies the I value is minimally influenced by the consistency increase.

▪ For relatively high values of consistency, as consistency increases, the I values decrease for all the pulps tested. This boundary is dependent on the fibre type.

The turbulence damping is higher in the case of the pine suspensions (longer and stiffer fibres), being lower for the recycled fibres suspension.

UNIVERSITY of COIMBRA

Page 25: Modelling Fiber Suspensions Flows Using Rheological Data

0

200

400

600

800

1000

1200

1400

1600

1800

2000

0 1 2 3 4 5 6 7

Velocity (m/s)

Pre

ssu

re D

rop

(P

a/m

)

Pine 0.76% - experimental

Pine 0.76% - predicted

Pine 1.0% - experimental

Pine 1.0% - predicted

005.0

0003.0%1

005.0

01.0%76.0

l

IC

l

IC

Results - Modelling

Comparison with experimental data

UNIVERSITY of COIMBRA

Page 26: Modelling Fiber Suspensions Flows Using Rheological Data

Results and Discussion

Simulated pressure drop for the flow of the recycled pulp suspension with 2.7% (w/w) consistency, at a velocity of 4.8 m/s.

Numerical Implementation

First, the numerical implementation was validated with water. Then, the pulp’s physical characteristics were introduced in the model.

UNIVERSITY of COIMBRA

Page 27: Modelling Fiber Suspensions Flows Using Rheological Data

Results and Discussion

Numerical ImplementationUNIVERSITY of COIMBRA

Kinetic energy profile for the recycled fibre suspension

0.72% consistency

Page 28: Modelling Fiber Suspensions Flows Using Rheological Data

Results and Discussion

Numerical ImplementationUNIVERSITY of COIMBRA

Kinetic energy profile for the recycled fibre suspension

2.7% consistency

Page 29: Modelling Fiber Suspensions Flows Using Rheological Data

Governing Equations

0 u

Fuupuut

u T

2

2

1 TT

k

T UUkUkt

k

k

CUUk

CUt

TT

T2

2

2

12

1

C 1C 2C k Constant

Value 0.09 1.44 1.92 1.0 1.3

Continuity equation

Conservation of momentum

Transport Equation for k

Transport Equation for ε

model constants:

UNIVERSITY of COIMBRA

Standard k-ε model

2kCT

Page 30: Modelling Fiber Suspensions Flows Using Rheological Data

Numerical Implementation

For the CFD modelling the Chemical Engineering module of COMSOL Multiphysics Software version 3.5 was used.

Geometry

4 m

The system to be modelled is basically a linear pipe (3 in diameter and 1 m long) where a pulp fibre suspension is flowing.

2D axial symmetry.

UNIVERSITY of COIMBRA

Page 31: Modelling Fiber Suspensions Flows Using Rheological Data

2D axial symmetry: mesh mode

In order to reach accurate results for the pressure drop, the mesh selected was a mapped mesh consisting of quadrilateral elements.

The mesh is more refined near the wall to resolve the viscous sublayer.

Numerical ImplementationUNIVERSITY of COIMBRA

Page 32: Modelling Fiber Suspensions Flows Using Rheological Data

Physics and Boundaries

Numerical Implementation

Inlet Boundary

Plug type cross section velocity profile;

The existence of particles, such as fibres, in a fluid flow induces a turbulence damping, thus the L and I values should be smaller then usually assumed for homogeneous fluids

Since the turbulent length scale is mentioned to be mainly dependent on the system geometry, it’s value was assumed to be constant for all the fibre types and all the consistencies. The intensity scale parameter was adjusted according to the pulp fibre type and concentration.

Outlet Boundary “Normal Stress, Normal Flow” function

Wall Logarithm wall function,

Symmetry Boundary

Axial Symmetry

UNIVERSITY of COIMBRA

Page 33: Modelling Fiber Suspensions Flows Using Rheological Data

Friction Factor

D

uC

L

Hf

2

2

Du

Re

.

RevsC f

Page 34: Modelling Fiber Suspensions Flows Using Rheological Data

Friction factor results

Friction Factor: Eucalypt pulp

Re

Cf

Page 35: Modelling Fiber Suspensions Flows Using Rheological Data

Friction Factor: distinct pulps

0.001

0.010

0.100

1.000

10.000

10 100 1000 10000 100000Re

Cf

recycled pulp suspension C=4.2% eucalypt pulp suspension C=3.5%eucalypt/pine pulp suspension C=3.2% pine pulp suspension C=3.4%water line

Friction factor results

Page 36: Modelling Fiber Suspensions Flows Using Rheological Data

Friction Factor

Correlations for not only as a function of Reynolds number, but also showing the dependence on consistency and pipe diameter for a range of pulp suspensions and for two different flow regimes were obtained;

fC

Page 37: Modelling Fiber Suspensions Flows Using Rheological Data

Rheological characterization

Typical rheogram for a pulp fibre suspension

UNIVERSITY of COIMBRA

Rate of shear, γ

Sh

ear

stre

ss,

δ

Newtonianfluid

τD

τY

Page 38: Modelling Fiber Suspensions Flows Using Rheological Data

Rheological characterization

Typical Rheograms

UNIVERSITY of COIMBRA

C=3.6%C=3.3 %C=2.9 %C=2.3 %

C=1.9%C=1.5 %C=1.0 %

Pine suspension

(Nm-2) ap (Pa.s)

0

50

100

150

200

250

300

350

400

0 2000 4000 6000 8000 10000

(1/s)

(N

/m2)

pine C=3.6%

pine C=3.3%

pine C=2.9%

pine C=2.3%

pine C=1.9%

pine C=1.5%

pine C=1.0%0.01

0.1

1

10

0 2000 4000 6000 8000 10000 12000

dV/dx (1/s)

pine C=3.6%

pine C=3.3%

pine C=2.9%

pine C=2.3%

pine C=1.9%

pine C=1.5%

pine C=1.0%