T -EEE TTT FFFI flEEE TTN FFFIF

.,. >..r

s

'.4'_'Eltr

,-tl.iÖ

--:8_,:=

/-

,:#

c)

a

C)ís

T -EEE ^TTT FFFI flEEE TTN FFFIF

ffiffilEIiÍE|*g

33 *n

GS trT

3n

ffiffilH$BiíittHit$lgH

EBBE NNN E8ilMB

ns*s -=- ^ECilAE

PARAMETERS ^{OF WHEY BY DTEFERENT} FILTRATION ^SYSTEMS

1,4,6,Ifniuersity _oÍ Szeged, Faculty of Engiruering, 6725 Szeged, Moszkoai'ryt: 5-7,

HÜNGáRY

z7,s Coroinus Llniaersity ^{of Budnpest,} Faculty of Food Scbnce,

HUNGARY

Absttact: Tfu ^largest quantities of by-Toduct ^{of the} dairy, namely ^{whey comes} ftom,tfu cheese-making' ^The whey proteins are usea"ay tn" agriculture"ii

animal'nutition,

^{and by the humnn nutrition as} .well;-dry soups, infant 7o,mula" a,d supplíments| ^T,lrc aim o7 ou| experiflents uas ^the separation of the lipid fraction ^of

uhty' Duing

^the

'measurments _o.bb

p*,

0,2 ym and 0.45 pm mictofiltration ^mtmbranes were used in aibrating ^membtane fltration.

"qrip*rrt ^{(VSE{) and} in a

laboratory tubular membrane modsle.

Duing

the mictofltratiory analytical ciwir"terisics, ^the fouling anit tfu rctenion ^{oalues were examined.} Itsing tfu VSEP and the htbular module made possible to corlpare ^the efftct of vibration, tle static mixer and/ t|e airflow on the separation.paÍaÍneteÍs.

'Ktyuortls:

_{cheese makiig,} experiments, lipid fractbn, laboratory tubular mtmbrane ^module

1.

INTRODUCTION

Liquid whey contflins

lactose,

aitamins,

^protein,

ilnd

rninerals, along utith traces ^of fftt.

During

^the

degreasing

Crearn

Can be

^skimmed

í,o*

^ahey.

ÍNhey Cream

is

tnore

salty, tangy, and

^,,Cheesy,, than ^{(,, srueet,,}

)

cream skimmed

í,o*

milk, find can be used to make zuLrcy butter. The other reason

of

tkw degrensin7

is

the

further

processin7 ^{oÍ the} ruhey

fw ^dry

^poT

der/nutitional supplement.

^The

meffibrane degreasing

metlPds are

^{nerLl ones {md}

tlrc

biggest

gap of it is the low flux ^and

^high

resistances. These

efficts could be ^mitigated

^by

used

meÍnbrfine

modes of us, i.e, static

mixer, aeration and aibrating.

Newtonian

fluids ^{such as} an

queous solution,

fire being turbulent flottt in ffiost industial

applications,

but roithin a

short pipe section ^this turbulence

is not

enough to equalize temperature

or concentration in-homogeneities. The use ,f

static

stirrers

^TDas made better aÍnal9affiation than

increase the

speed

or the pressure during

^t|rc

process. The

flux ^is

^increased

^and

the operating cost

is

decreased

at

tubular rnembrflnes

nith

static

mixer ^(Krstic et fll, 2a04. Similar result

^TÍ)as obtained

uith an

alternqtiae design equiprnent ^to

1,

Szilárd szÉrcaL,2, ^Zsuzs

^mrfi

^K]HANY,3,

^Eszter

F)GARÁSSY'

+.

!őzsefcsÁNÁ

^{DI, s.}

Gyula VATAI,

^{6, Cecilia}

H)DUR

ASSAYING OF THE FILTRATION

increase turbulence

and

other type

of

membranes

as u)ell

(Bellhouse

et al. 20AL Costigan et

^aI.

2002). The

fouling

^of tlrc membranes u)ns possible to decrease t tlrc introduction of gas into the

liquid

(Laboire et

al.

1998, Cabassud et aI. 2001,

Cui

and

Wright,

1996). The

introduction

of a specific gas ^-

in this

case

sir - directly into

^the

fluid

^{created a}

tr.uo-phase

gasfliquid flow. ^The efficiency of

the

sepnration is influenced by thl position of

the membrÍune (aerticfll or harizontal) gnd ^the direction of the floTl, (up or doutn).

The aerütion method is ^limited by ^the,

_9ils

distribution and

the manfiSernent

of this

process

(Derradji, 20Aq, Duing the uibratory

^shear

enhanced process

(VSEP),

the

filtering

^parameters (flu*, retention

and

resistances) ^Tuere inztestigated

by the

effects

of the uibration. This is

another

solution

to decrease

fouling

(Frappart et aI. 2008,

Hod r et al, 2aB). The

shears strengths

at

thÍ surface oÍ the Ínembrane Can be increased by

aary

tke frequency

of

the

aibratory

mpmbrane module.

The

polaization layer,

the resistance ztalues, and

the fouling

^íL}ere

^measured by the effect ,Í

aibration, and the euolution of retention

ualues

Zlere

me

sured by ^{the effect} ,Í ^increase ,f

Á'rrÁ

^Tffitr fl&ffiCA CíÖ R

WNjrEhrsrls

- ^BnÍletfu

^of

Etlgilrcerillg

Tonre \rII [2tr14tr aibrational amplitude (Ahmadu et

al.

2009,

Hod r

et al. 20a9, Kertész et al, 2a10), 2,

MATERIAL9 ÁND METHODS

Szueet cheese

uthey

?.ofis used

for

measurement zohich came

frr* Soma Budapest Ltd. Its

basic analytical parameters ^are:

fat

content: 0.18 m/m%,

protein

content: 0.33 m/m%,

milk sugar

content:

2.61 m/m%, dry mateials: 3.72 m/m%,

total

protein content: A.47 m/m%. The

^degreasing

process utas made

by

membrane separation. These basic parameters tuere rneasured

by Bentley milk

analyzer equipment.

The air injection and/or static steeing

method rnere

implemented by tubular and hollow fiber

membranes

with

0.45,

0.2 microns,

0.05 microns

cut

off aalue. The

tubular

membrane was 250 mm Iength,

and it

rpas

included 1

tube uthich has an

internal

diameter

oÍ 7 ,nm.. The applied

static mixer TL)as fi 250 mm lenght

Helix

_Upe metal static

stirrer with

a pitch of 0,0A6 m and an

inner

radius

of

the

stirrer of

0.A035

m. (KenicsTM, Helix)

as

illustrated in Figure

1.. The

KenicsTM

type static

stirrer

^(made

by plastic material)

rL,as used also ntith a length of 2a1 ffiffi, and a thickness of 1 ffiffi, a diameter of 6.35 rnm, The

flux

^ntas performed on

100 L/h recirculation Jloutrate, on 0.2 MPa transmembrane pressures and on 2A L/h air injection rate. In all meilsurments the initial amount of feed

^TI)íts

2 L of sweet

rtthey. The ternperature

TDas

perTnílnent

sa"C

^degree

during

the tests. The airfloztt aws introduced

inta

^the

fluid

Jloro before ^the membrane module.

Figure 1: The Kenics

TM

(right) and

the Helix type static stirrers

Vibratory fiItration ^equipment set

marketed by Nezu Logic International

Crrp.

and this equipment Toas used

at L-mode (L: laboratory

methods: thp module comprises one disk-shaped membrane

aith

an actiae

filter

^surface 503cm2). The

VSEP

systent consists of disk-shaped flat-sheet membranes. This laboratory module attached to a central shaft. The shaft ntas rotated a short distance at ^a frequency

of

54

Hz.

0.2

lm

^{cut-off aalues membranes (mad,}

,Í

polyethersulfone) TÍ1ere used during

^the

ftrcasurements,

on a

transmembrane pressure at

A3 MPa. In

this equipment the

initial

amount

of

feed was ^1.0 L of sweet zuhey.

The samples were taken at different

interaals

during

the measurement

from

^retentate

^and

^also

í,,*

^permeate.

3, rdiEgULTS

ÁND DIsCUssIoN

The retention

of

the

fat

^{component zuas important}

in our

research prograrnme,

and

^TL,e Tt)ere able to

retain

more than 50%

at

lonr pressure utith

using static mixer.

The retention aalues were measured

at dffirent transmembrane pressures

and recirculation floztt rates by

,

0.45 micron pore size membrane (Fig. 2). The main objectiue utas to hold back the

fat

^{molecules fis much as possible}

^and

^to

grrc

up the other molecules.

ltrifil l.ctia| l}rohein ,-x:%.ffi1gf'É*p''1ffi

tl..tll]rt.sc|Íds ^.E]'idÉ*rf.ír*+Í*i f ^*f F:fP1 ^l..l

2 trar 5flLlh

;l 1 bor írfl"]h n 2 bar ^1l00Uh

1 bar il00llh x 2.bar í5{}Líh s 1 t*{r 150LJh

|3.,ií l-lct'cse

illfri Protiein

i{ldi ltE

!..n.i.f &,v.í1:g'iii..&H*w!.-.ffi ..:'.i

ffi íi(fi}aiÍ*'ie.6*í'ffi's'$E

{í';Jff !fu !li$,*.{F.ie'aiáff;vTf '4il t$:1.*';...,1

ffi

Ü10203Ü405060

Figure 2: The retention aalues (R) of the

dffirent

components on 0.45 ym cut affaalue tubular membrane

Th,

figrre

^{shouts that the}

fat

^{micelles zrtere retained}

the greatest amount r.phen A.1 MPa

transmembrane pressure

and

15A

Llh

recirculation

Jloro rate

^T,Üere

used, The other

component TL)as

rejected

at

highest leztel

at

these parameters also.

But

these higLrcst

aalue itself is not

enough

for

- ^Bulnethl

^tlf

Ellgíneer.illg

degreasing aspects since

it is

less than 50%, sa the 0,45 micrometers pore size could be said to big

for

this task.

Since the work

u)as

continues roith 0.2

microns

pore size tubular

membranes.

Better

retention r.talues Tnere measured

at lou:er

transmembrane

pressure

^(0.1

MPa) at the 0.2 micron pore

size

membrane than at the A.45 micron pore

^size membrane.

The bigger

transmembrane pressure

resulted

^rL,orse

retention

ztalues;

this tendenry

is the same

at this

pore

size as well. The Figure

³

presents the ,ffi,t ,Í ^diffirent

^process

rrangements

on

the

retention.

ThB

static stirrer

has

a

decreasing effect

on

the retention of protein, Iactose and

dry

matter.

Only

^the

fat ,rrrr\isn

^utas increased

by using the static stirrer;

^maximum

ualue was tneítsured at

TMP:

0,2

MPa,

q,: 1"00 L,h.

Tfu

tubular membrane gflae better retention aalues than the

capillrry

membranes, but as our goal

is

to

minimise

the

fat ^{content and to}

^keep

^the

^otlrcr

ingredients it is clearly be

seemed

Í,o*

^the

fttaasurements the 0.2

microns tubular

^membrane

with static stirrer is

the best

solution among

^the applied arrfingements.

Bi{r lotal t3rotei

B:íi l3t"x'ry|i.i;i

Bl{r l-aqto:t

Itr1F'r Ptotein

Bli Fa[

duing the normal filtration

^process,

and with

combined the

air injection,

the

flux

^{was decreased}

slight by the air

flow

^{on a 0.2 pm membrane. When}

the Helix static mixer

element n)as

used in

^the

fiItration

^{process, tLrc}

flux

^{aalues rnere increased}

tuto times greater

extent,

frr* ^{L8 l/mzh to}

⁴⁰

Um2h uthen the

Helix static mixer

roas used

with

air injection.

The increase

of

the

flux ^{is holding} until

the 0.2

MPa

transmembrane pressures; because on higher pressure ualues

this

increase

of

the

flux ^is

^started

to show a strong deceleration (Figure 4).

IMen

thr

air injection process was used alone,

the

flux aalues remained aery low; therefore the air injection method itself is ^not a

recornmended method

for

^uthey processing.INhen the

Helix

static

mixer

was used alone

in

^tLrc equipment

under

^the

same

parameters, the

flux ^aalues

^{Tt)ere showed}

higher aalues than the expeiments ztrith air injection, but aboae 0.2 MPa

transmembrane pressures ualues, the

flux

^zttas

^strongly

^decreased,

therefore it u)as justifed to use the

^loruer

tr ansmemb r ane pr e s sur ^{e .}

18

z 0

n ffi* W"

%

iiá*Hllk ffi,ee...

ffi

*ffi$lii*ars

ffi

ffi

16 L4 12

*- = ₁₀

E i**r Jó

&E*

ffi

KBl 2.bal '0Ü0Líh ', KM 2 bar tMLJh

a KM 2 bar t000Llh CSM+$K 2 bar Í00L]h r CSM+SK 2 bar trÜ0LJh 'l Cssi+5K 2 bar'!Ü0ljh

[sM 2 bar íOÜLih e CS&'l2 bar'!00LJh r CSM 2 bar í*0Lih

CSM 1 bar !Ü0Llh r csM ^{.l bar} í ^ÜLjh

r CSM 1 bar trÜOlih

irlii*#i*Jaiir 0.25

ffi

0x0203Ü405060

R{%}

Figure 3: The retentian aalues (R) of the different CoÍnponents were measured in 0.2 pm cut offaalue tubular membrane with (CSM+SK) and without static

stirrer (KM-capillary membrane,

CSM -

^tubular

membrane)

The ÍIu* ^aalues

^T,zere

^measured at 0.2 MPa transmembrane pressure, and at rcO

^L/h

recirculation Ílo* ^{rate by} fl 0,45 rm ^tubular

membrane.The

ÍIux

aalues are started at 6a l/mzh.

The

flux

^ualues u)ere s\pTtted 17-18

Um'h

^aalues

0.00 0.05

0.10 0.15

^0.20

apil _[MPa]

Figure 4: Tfu changes of whey ftux ^(l)

as a function ^of trasnmembrane pressure at

dffirent

re cir culation flow ^{r ate}

The 0.2

microns

membrane u)ils used ruith Kenics type

static mixer in

second period of

our

research prograrumes. The

flux

^aalues

(l

=

4S Um2h)

^ll)ere

increased by the Kenics static mixer, but

the increasing zzas not as high as

using

the

Helix-type

static

mixer (l

= 53

Umzh), Our

experiments znere

continued by a 0.05 microns pore size

tubular membrane to

compaing

the receiued data

with

the other

tubular

membranes different dnta. The

flux

Ácffi& TnÍdhírcÁ

^Ce}R

1ryj\ilr'Nsís

- tsulletill

^ttf

E'llgilleer,ílls

T,t"mte WX

[?W$

u:as increased 50-80%

during

the

fiItration

^process

by ^using a 0.AS microns pore size

tubular membrane zpith

a Kinetics static mixer.

^The

flux

was ^{decreased strongly after} 0.3 MPa transmembrane pressures. Tfu uibratory

shear enhanced membrane

filtration

^TL,as examined by a 0.2

p*

pore size

microfltration

membrane, on 4.3

MPa transmembrane pressures zuith using oibration at 54 Hz aibrational frequenry

^and

w i thou t u sing a :b_ry t io1,{_Fj gyfr_,S_)_, r20.00

t

lncreased pressure ⁺ 54Hz vibr.

ffi Increased pressure ⁺ OHz vibr.

100.00

I

*Í&

where due to their sizes ^(3.5 miuons),

^these

molecules were fouled inside the

membrane

capillaies. The increased retention ualues

and their absolute magnitude utere been

significant

by

fot ^{content, Tlrc retention aalues} ,Í ^the

^small

coffiponents uere increasedby ^the fouled ^pores.

iRetentbn values

:;;l

^[Yoll

m 0.2 mm

54ln t

^{0.2 mm}

0Hz

80.00

S

(! ^oo.oo

3

40.00

i -=- --- __

0.00 fat

0.0 0.1 0.2 0.3

^4.4

TMP (MPa)

Figure 5: The changes of whey

flux

^(l)

as a function ^of trasnmembrane pressure at uibrated ⁽⁵⁴ Hz) and non aibrated methods

The retention aalues ^'u)ere mensured only

from

^the

fat

molecules.

Tlrc examination of the

resistance

aalues was

shorued

that the gel layer and

^the

membrane resistance aalues showed the

^sftme

magnitude aalues. The fouling resistance

^u)ns

shoroed an order of magnitude lortter aalue than the tzoo other determinatiae resistance aalues ^before.

In non-aibrating mode, not only the

total resistance

oalue

^T.D

s

showed differences,

but its structure

and

distribution

as utell.

Without using aibration during

the separation process, ^the

flux aalues were

^shozued

fow times lower; the

total

resistance aalue was shoarcd one ordercf

magnitude higher; and the fouling

^resistance

ualues

rL,ere

shorued tuo orders ,f

^magnitude

higker aalues.

Thc drag

resistance aalues

uere

decreased

by

the aibration, therefore this

clunge

^u)as allowed ^the

fat molecules to

^moT)e

and accumulate on

^the

membrane surface (Figure 6)

^T'he

flexible fot

molecules u)ere moued

into the capillaies of

^the

membrane

under

pressure

and

roithout aibration,

.- ^.- - -t- -, , .t

protein , . -.--r-

_.--.----/

Ectose

dry materials

Figure 6: The retention aalue of most important components of whey

4. CONCTUSION

The expeiences shouted that the 0.45 microns pore

size

membrane

could slightly hold back

the

fat

molecules, due to tlrcir larger pore size. The desired

filtration results ^{rrlere achieaed} by

^the

measurements of A.2

MPa

and 100 L/h.

The

45 %

higÍter

ÍIu* ^aalues

tllere measured by

Helix static stirrer against the

normal

fltration

process,

but

the combination

of

the

static

stircer

and

the

air injection

were made the highest

flux

aalues (30 % higher thnn the

filtration

process by the

Helix

static

stirrer)

under the same conditions.

Comparing

the tzoo

dffirent static stirrers, it

was

found, that the 15 % higher flu*

^{ztalues were}

measured

by Helix static stirrer than

the Kenics

stirrer. This

means

that

the separation

oÍ

the

fat

content utas easier and more effectiae by

using

^the combination of static stirrer and air iniection.

The

aibratory

shear enttnnced process rl,frs shouted that not

only

the retention ualues of ^the

fat

^content Tzere increased

utithout aibration, but the

other elements retention

aalues

too,

30a

% higher

Íl,*

aalues

znere measured

by 5a Hz aibration

thnn utithout aibration.

-

Bun:letilt of E"ttgitte.erurug

ACKNOWLEDGEMENTS

This research was supported by the European

Union and the State of Hungary, co-financed ?V

^the

European

Social

Eund

in

^the _framework ^of

TAMOP- 4.2.4.A/

2-11f1-2012-0001

'National

Excellence

Program'.

*

The authors

acknowledge

the

contribution

,f

^the

MEMFIDAL prograrn (EUREKA HU

^08-7-2010-

\An)

and the

OTKA

K105A21program.

REFERENCES

tL,l

Ahmadun F.-R., Pendnshteh A., Abdullah L. C.,

Biak D. R. A.,

Madaeni

S. 5., Abidin Z.

Z.

(2009), Reaiew

of

technologies

for oil

^{and gas}

produced water treatment. lournal ^{of Hazardous}

Mateials

^{VoI.: 170.} pp. 530-551

tz.l Bellhouse,8.1., Costigan, G,

^{Abhinaua, K.,}

Merry, A.

^{(2001), The} performance

of

kelical screwthread

inserts in tubular

^membranes.

Separation and Purifcation Technology Vol.: 22- 23,pp. 89-113

t3.l

^{Cabassud, C.,}

^Laboie,

^5., Durand-Bourlier, L.,

Lainé, I.M. ^(2001), Air sparging

in ultrafiItration hollow fb ^{ers : r} elationship betraeen

ÍI,*

enhancement,

cake

characteristics and hydrodynamic parameters,

I. ^Membr.

^Science,

Vol.: 18L, pp57-69.

t4.l

^{Costigan, G.,} Bellhouse, 8.1., Picard,

C.

^(2002),

Flux enhancement in microfiltration

^by

corkscrew aortices formed in helical

_Íto*

pfissages, lournal ^of Membrane Science, Vol.:246 pp.179-188.

ts.l ^{Cui, 2,F., Wright, K.LT. (1996),}

^Flux

enhnncements with gas sparging

in

downwards

crossflow ultrafltration:

performances ^and mechanisms,

I.

^{Membr. Science, Vol.: 1L7, pp.}

L09-116.

t6.l

^Derradji,

^A.F.,

Bernabeu-Madico,

A.,

Taha, 5., Dorange,

G.

_QA00), ^-fhe

,ff rt

of a static mixer

oft the ultrafiItration ,Í a

^two-phase

_ÍIo,,

Desalination, Vol.: ^{128, pP.} 223-nA.

t7.l

^Frappart

^M,

^Iaffrin

^M.

^Y.,

^Ding

^L.

^H.,

^Espina

V.

QA08), Effect

of

ribration ftequenry ^and membrane shear rate on nanofiltration of diluted

milk, using a aibratory dynamic fltration

system. Separation and Purifcation Technology, Vol.: 62,W.212-221.

t8.] ^{Hod r, C.}

^Kertész,

^Sz.,

^Csanádi

^|.,

^{Szabő G,,}

Lászlő

Zs.

(2009), Inaestigation

of

Vibratory- shrar Enhanced Processing System. Progress in

Agricultural

Engineering Sciences, Vol.: 5, pp.

97-1_1_0.

t9,] ^Hod

^r C,, Kertész

Sz,

Szép A., Keszthetyi-Szabő

G,,

Lászlő

Zs.

_Qa13), Modeling of Membrane Separation and

Applying

Combined Operations

at Biosystems, Progress in

Agricultural Engineering Sciences VoL 9:(1) pp. 3-25.

t10.]

^Kertész

Sz, Erbasi E,, Lászlő Zs,,

Hoaorka- Horaáth

Zs.,

Szabő G.,

Hod r _{C. Qffia)} oily

rnasteusaters separation

by

ultrafiltration.

IWA Regional Conference and Exhibition

^on

Membrane

Technology

a ^Wafur

^Reuse

_í"ll

paper, 18-22 October 2010 Istanbul-Turkey, 351- 355.

F1,] ^Krstié, D,M.,

Tekié.,

M.N.,

CariÓ,, M.D,, MilanoaiÓ,

s, D.

^(2002), The

,ff,,t

of tubulence promoter

on

cross-flow microfiltration

of

skim milk, lournal ^of Membrane Science, VoL: 208 pp.

303-314.

[12.] ^Laborie, 5., Cabassud,

C,

Durand-Bourlier, L.,

Lainé, I.M.

^(1998),

Fouling control w

^air

sparging inside hollow fibre ^membranes

-

^effect

on energy consuruption, Desalination, Vol.: ^118, pp. 189-196.

ffi{JB_ttT[N

^#F"

_WLHffi

copyright @

University "I'OLITEHNICA" Timisoara, Faculty of Engineering Hunedoara,

5, Revolutiei,

331128, Hunedoara,

ROMANIA

http:/ _/ ac ta. ^f ih.upt.ro

pl

.,#

= ,J':i, lpfi6

:.1t

!l! ,it

,Jrf

lfí{Trt!k|"l'{.'T'E:"j*. ^.:

. !r {;

5n ry

*$*.r''L *ffiH| trFppÍi

tilKtr

iilEil ^$rrFp.q$n

ps 3E

_3&

35

,$s

í*td,&ds*

gÜ[

_5*'&5m$

3r;*{dlq*

finn

*'&.$*