applied
sciences
Re iew
Elec ical Tomog aphy: A Re iew o Con igu a ions,
and Applica ion o Fib e Flow
Suspensions Cha ac e isa ion
Ped o Faia 1,* , Rui Sil a 2, Ma ia G. Ras ei o 2and Fe nando Ga cia 2
1Depa men o Elec ic and Compu e s Enginee ing, Facul y o Sciences and Technology, CEMMPRE,
Uni e si y o Coimb a, 3030-290 Coimb a, Po ugal
2
Depa men o Chemical Enginee ing, Facul y o Sciences and Technology, CIEPQPF, Uni e si y o Coimb a,
3030-790 Coimb a, Po ugal; [email p o ec ed] (R.S.); [email p o ec ed] (M.G.R.); ga [email p o ec ed] (F.G.)
*Co espondence: [email p o ec ed]; Tel.: +351-239-796-200
Recei ed: 6 Ma ch 2020; Accep ed: 19 Ma ch 2020; Published: 30 Ma ch 2020


Fea u ed Applica ion: The depic ed omog aphic echnique, Elec ical Tomog aphy, is a
non-in asi e and emo ely ope a ed imaging sys em ha makes i possible o ob ain aluable
in o ma ion abou mul iphase sys em low beha iou s: The 2D/3D images ob ained a e o
ex eme impo ance no only in he de elopmen o p edic ion models, bu also in hei eal- ime
moni o ing and con ol. I has high applicabili y in se e al indus ial en i onmen s, such as
chemis y, pape , pe oleum, wa e and was es indus ies, jus o men ion some.
Abs ac :
Unde s anding he beha iou o suspension lows con inues o be a subjec o g ea in e es
conside ing i s indus ial ele ance, ega dless o he long ime and e o dedica ed o i by he
scien i ic and indus ial communi ies. In o ma ion abou se e al low cha ac e is ics, such as low
egimen, ela i e eloci y be ween phases, and spa ial dis ibu ion o he phases, a e essen ial o he
de elopmen o exac models o desc ip ion o p ocesses in ol ing pulp suspension. Among he
di e se non-in asi e echniques o low cha ac e isa ion ha ha e been epo ed in he li e a u e o
ob aining expe imen al da a abou suspension low in di e en p ocesses, Elec ical Tomog aphy is
one o he mos in e es ing, since i p esen s pe haps he bes comp omise among cos , po abili y,
and, abo e all, sa e y o handling (indeed he e is no need o use adia ion, which equi es special
ca e when using i ). In his pape , a b ie e iew and compa ison be ween exis ing echnologies o
pulp suspension low moni o ing will be p esen ed, oge he wi h hei s eng hs and weaknesses.
Emphasis is gi en o Elec ical Tomog aphy, because i o e s he abo e-men ioned comp omise and
hus was he s a egy adop ed by he au ho s o cha ac e ise di e en low p ocesses (solid–liquid,
liquid–liquid, ib es, e c.). The p oduced po able EIT sys em is desc ibed, and examples o esul s o
i s use o pulp suspension low cha ac e isa ion a e epo ed and discussed.
Keywo ds:
ib e suspensions low; elec ical omog aphy; complex lows; dispe se phase dis ibu ion;
eloci y e alua ion
1. In oduc ion
Fo he ob en ion o omog aphic in o ma ion, di e en measu emen me hods can be ollowed
wi h espec o he sensing p ocedu e: in usi e, non-in usi e, in asi e, and non-in asi e. The choice
o an app oach is he esul o a wise choice made by he use s. By ‘in asi e’ is mean ha he senso
su ace is pa allel o he inne zone o he wall o he channel unde s udy, while he e m ‘in usi e’
is used when he senso i sel in e ac s wi h he low being isualised. In opposi ion non-in asi e
e e s o cases in which he senso is loca ed on he ou side o he wall, wi hou c ossing i , while
Appl. Sci. 2020,10, 2355; doi:10.3390/app10072355 www.mdpi.com/jou nal/applsci
Appl. Sci. 2020,10, 2355 2 o 30
non-in usi e indica es cases in which he senso placemen o ces he pipe wall o be c ossed, bu
wi hou dis u bing he low unde inspec ion [
1
]. In ecen yea s, non-in asi e p ocedu es ha e shown
a high deg ee o po en ial g ow h in many a eas, such as in he chemical indus y. When compa ed wi h
non-in usi e p ocedu es, non-in asi e app oaches b ing bene i s o he abo e-men ioned indus y
sec o s, such as educ ion o haza ds due o ope a ion o co osi e o adioac i e ma e ials, inc ease in
ope a o sa e y, e c. Consequen ly, he non-in asi e app oach is he imaging s a egy mos commonly
used in p ocess plan s.
Fo ha d- ield omog aphy, he case mus be conside ed in which he e he measu ed pa ame e
dis ibu ion and he sensi i i y o he imaged media in he en i e 3D domain a e independen o
one ano he [
2
]; op ical, ul asonic and X- ay a e examples o his ype o omog aphy. On he o he
hand, so - ield omog aphy e e s o cases in which a dependency is ound, o e he en i e olume,
be ween he medium sensi i i y and he dis ibu ion o he measu es o he desi ed a iable [
3
].
Elec ical omog aphy is he mos ele an membe o his class, bu i p esen s an impo an e o
ega ding he esolu ion o he in e se p oblem on he domain unde obse a ion.
2. Pulp Suspensions Imaging O e iew
Pulp suspensions a e cha ac e ised as being opaque and in ica e mix u es ha exhibi a qui e
he e ogeneous dis ibu ion o ib es unde low, p esen ing a eas in which locs a e p esen ( he e m
locs e e s o a high concen a ion o ib es). In addi ion, along he en i e pulp suspension and abo e
ce ain concen a ions, as a unc ion o he obse ed shea a es, a con inuous ib e ne wo k o ms.
Fo he pu poses o p ocess con ol nowadays, knowledge ega ding low cha ac e is ics is essen ial,
once he suspensions ha e been subjec ed o high shea o ces. The numbe s ega ding u no e in
he pulp and pape indus y con i m he impo ance o his knowledge, as in some coun ies hey
exceed he e u n om o he indus ial a eas, as is he case in Sweden and Canada. Unde s anding
and con olling he manu ac u ing p ocess enables he de elopmen o newe p oduc s wi h a high
inno a ion con en . Addi ionally, ob aining p ecise alues o he eloci y p o iles associa ed wi h
a ce ain shea a e in a pulp suspension is also o high impo ance o op imising he low and,
consequen ly, he ene gy consump ion; o he la e , howe e , he e is sca ce in o ma ion epo ed in
he li e a u e.
Di e se low imaging me hods, along wi h pa icle ollowing elocime y app oaches, ha e
been desc ibed in he li e a u e; howe e , mos o hese need o be expe imen ally ained, wi h
da a acquisi ion p ocedu es hus usually las ing qui e a long ime, and hey can ypically only be
used in non-opaque suspensions. Consequen ly, i is no possible o use hem in pulp suspension
lows no belonging o he dilu e ange. Fo da a collec ion in pulp suspensions, among he mo e
p omising echniques can be ound Lase Dopple Anemome y (LDA) (Op ome GmbH, Da ms ad ,
Ge many) [
4
], Magne ic Resonance Imaging (MRI) (Na ional Ins i u e o Biomedical Imaging and
Bioenginee ing, Be hesda, MD, USA) [
5
], Ul asonic Veloci y P o iling (UVP) (MET-FLOW S.A.,
Lausanne, Swi ze land) [
6
], Op ical Cohe ence Tomog aphy (OCT) (PHOENIX TECHNOLOGY
GROUP,Pleasan on, CA,USA)[
7
], Pa icleImageVelocime y(PIV)(DANTECDYNAMICS,Sko lunde,
Denma k) [
8
], X- ay omog aphy (Depa men o Physics P.O., Uni e si y o Jy äskylä, Jy äskylän
yliopis o, Finland) [
9
], Gamma ay densi ome y [
10
], and Elec ical Tomog aphy (ET) (PASI SRL,
To ino, I aly) [11].
Magne ic Resonance Imaging is a omog aphic echnology wi h a high spa ial esolu ion ha can
supply in o ma ion ega ding he beha iou o he p o ons p esen in a sys em, which a e ypically
con ained in he 1H nuclei o wa e . MRI can be used, among o he applica ions, o image he mo ion
o wa e , e.g., exp essed in he o m o eloci y p o iles. The majo limi a ions o MRI a e he ypes o
pa icles and he size o he sys em ha can be s udied, oge he wi h he size, weigh and cos o he
ins umen a ion. Only pa icles con aining MR-sensi i e nuclei, such as 1H, can be de ec ed [
12
,
13
].
The maximum diame e o he sys em is bounded by he inne diame e o he coil in he magne .
Thus, MRI expe imen s a e ypically limi ed o labo a o y-scale luidised beds o pipes wi h small
Appl. Sci. 2020,10, 2355 3 o 30
diame e s [
14
]. In summa y, he s eng h o MRI is in s udying cen ime e-scale sys ems a spa ial
esolu ions o ~100
µ
m. Time-a e aged eloci y p o iles o dilu ed pulp suspension ha e been ob ained
using MRI wi h ib e concen a ions o up o 0.86% (w/w) [
15
–
20
]. Howe e , when compa ed wi h
LDA and UPV, MRI p esen s some disad an ages, such as longe obse a ion imes and equipmen
cos . A ola e al. [
20
] we e able, o ga he in o ma ion abou he suspension eloci y p o iles o 0.5%
ib es dilu ed in wa e (w/w) using MRI in a ma e o milliseconds. Also using MRI, unde s eady
low condi ions, and o pulp suspensions wi h concen a ions abo e 3% (w/w), Seymou e al. [
21
]
we e able o measu e ime-a e aged eloci y p o iles. Despi e hese lab ials ha ha e aken place
a ound he wo ld, wide MRI usage in indus y is no expec ed in he coming yea s.
Lase Dopple Anemome y (LDA) is an op ical echnique used o in es iga e eloci y and
u bulence in gas, liquid, and mixed luids, among o he applica ions [
22
]. The basic idea unde lying
LDA is o measu e he eloci y o iny pa icles anspo ed by he low. I hese pa icles a e small
enough, hei eloci y is assumed o be ha o he liquid s eam, and LDA hus p o ides a measu e o
he local ins an aneous eloci y, he mean eloci y, and he u bulen quan i ies. Lase anemome e s
o e unique ad an ages in compa ison wi h o he luid low ins umen a ion [
22
]: (1) Non-con ac
op ical measu emen —LDA p obes he low wi h ocused lase beams and can de e mine he eloci y
wi hou dis u bing he low; (2) Well-de ined di ec ional esponse— he quan i y measu ed by LDA is
he p ojec ion o he eloci y ec o in he measu ing di ec ion de ined by he op ical sys em; (3) High
spa ial and empo al esolu ion— he op ics o he lase anemome e a e able o de ine a e y small
measu ing olume, hus p o iding good spa ial esolu ion and allowing o local measu emen o
eloci y; (4) Mul i-componen and mul i-di ec ional measu emen s—combina ions o lase anemome e
sys ems wi h componen sepa a ion based on colou , pola isa ion o equency shi allow one-, wo-,
o h ee-componen LDA sys ems o be pu oge he based on common op ical modules. So indeed,
LDA has a lo o ad an ages, bu some comp omise has o be made when selec ing and se ing up lase
anemome e sys ems; indeed, among he disad an ages o LDA a e he high cos o he equipmen ,
he need o a anspa en low h ough which he ligh beams can pass (usually co esponding o
a low concen a ion o he dispe se phase), and he ac ha hey do no gi e con inuous eloci y
signals. Ne e heless, he special p ope ies o he gas lase s (such as high ene gy, spa ial and empo al
cohe ence, and s abili y) make his me hod applicable o sol ing a la ge numbe o p oblems. LDA has
been used o examine low concen a ed pulp suspensions con aining ai bulbs by Ek e al. [
23
]; ca y
ou a s udy o 0.5% ib e suspensions in wa e (w/w) by Ke ekes e al. [
24
]; s udy pulp suspension low
a concen a ions be ween 1.2 and 12 g pe li e o wa e , using anspa en model sys ems, based on
he e ac i e index ma ching o he liquid and he ib es by S een [
25
,
26
]; and, using he same model
sys em p esen ed by S een, pe o ming measu emen s in a s i ed ank wi h ib e concen a ions in he
ange 3–20% (w/w) by bo h Ande sson e al. [27] and by Pe e sson and Rasmussen [28,29].
Since Acous ic Dopple de ices achie ed easy handling and low cos , hei use has expe ienced a
high a e o g ow h in ecen yea s. Dopple e ec -based echog aphy equipmen allows he ab ica ion
o sys ems ha a e able o measu e eloci y p o iles nea ly ins an aneously [
30
]. The Ul asonic
Veloci y P o ile (UVP) echnique was ini ially limi ed o he inspec ion o non- anspa en luids [
31
],
and o measu emen s h ough pipe walls in small-scale sys ems [
32
]. In he UVP echnique, a p obe,
ac ing i s ly as an emi e , p oduces an ul asound signal ha a els along a ce ain di ec ion, usually
he suspension di ec ion, and la e , ac ing as a ecei e , senses he e lexions o igina ed by pa icles
p esen in he suspension; he ime, as well as he Dopple equency shi be ween emission and
e lec ions, is measu ed. Using he alue o sound eloci y in he luid, he dis ance and eloci y o
he pa icles o igina ing he echoes ( he e lexions), along he suspension di ec ion, a e calcula ed.
UVP was o iginally es ablished o 1D measu emen s; howe e , p omising 2D eloci y p o iles
ha e ecen ly been published [
30
]. The UVP echnique has been used ex ensi ely in ecen yea s o
ob ain ins an aneous eloci y p o iles, bu e y ew s udies on pulp suspensions can be ound in
he li e a u e. Hi simäki [
33
] used an ea ly ul asonic p o iling echnique o ob ain adial eloci y
p o iles in pulp suspensions wi h concen a ions up o 1% (w/w). Ka ema and his co-wo ke s [
34
,
35
]
Appl. Sci. 2020,10, 2355 4 o 30
cha ac e ised eloci y luc ua ions and s udied pape o ma ion by luidisa ion and e loccula ion o
wood pulp suspensions wi h concen a ions up o 1% (w/w) using UVP. Wiklund e al. [
31
,
36
] ob ained
ins an aneous eloci y p o iles in a s eady lamina low o 0.5–3% (w/w) and la e up o 7.8% (w/w)
cellulose pulp suspensions using UVP. Claesson and he co-wo ke s [
37
] used UVP o cha ac e ise
pulp ib e suspensions h ough a sudden expansion a wo di e en concen a ions, 1.8% and 2.8%
(w/w), and a di e se eloci ies, om 1 up o 2.2 m/s: They we e able o ins an aneously ob ain eloci y
p o iles as well as axial eal mean squa e eloci y luc ua ions in he e alua ed semi-concen a ed lows
jus a e he expansion. To enable imp o ed low in o ma ion o highe concen a ions, Ko z
é
e
al. [
38
] es ed a pulsed elocime y-based sys em o measu e eloci y p o iles o cellulose lows o wo
di e en concen a ions, 6.9% and 8% (w/w), in h ee di e en pipe diame e s (16, 22.5 and 52.8 mm).
Also using a pulsed app oach, Hanjiang and colleagues [
39
] ob ained eloci y p o iles o ib e lows in
a ec angula channel, and in es iga ed he in luence o ib e concen a ion and Reynolds numbe on
he shape o he measu ed p o iles; hey used na u al wood ib es (wi h an a e age leng h and diame e
o 2.3 and 35
µ
m, espec i ely), and he es s we e conduc ed wi h a ying a e age low eloci ies o
be ween 0.16 and 7.0 m/s, and di e en ib es concen a ions be ween 0.05% and 1.0% (w/w).
Op ical Cohe ence Tomog aphy (OCT) is a ecen echnique ha allows he in e nal s uc u e and
he mo ion beha iou o opaque suspensions o be measu ed simul aneously and wi h high spa ial
esolu ion. Depending on he echnology used in he OCT cons uc ion, he axial scan a es can go up
o hund eds o kHz [
7
]. OCT ope a ing modus is qui e simila o ul asound imaging, wi h he main
di e ence being i s ope a ing equency; while OCT ope a es in he ligh equency band, ul asound
uses he sound band. The dep h o he ob ained images depends o a la ge ex en on he op ical
p ope ies o he domain; consequen ly, i s dis ance om he wall can a y be ween mic ome es and
some millime es [
40
]. This echnique was used by Ka aja e al. [
41
] o cha ac e ise he bounda y laye
o mic o- ib illa ed cellulose lowing in a s aigh pipe, o consis encies in he ange 0.4–1.6% (w/w).
Howe e , in he case o ul asound sys ems, which usually ope a e in he MHz equency band, he
wa es a e eadily ansmi ed o he domain, making i easy o ob ain images o signi ican dep h:
ne e heless, a hose equencies hey can also su e om high a enua ion, limi ing hen he used
equencies. Fu he mo e, i is much mo e di icul o ocus sound wa es han ligh wa es [42].
Pa icle Image Velocime y (PIV) is a lase measu ing echnique, as is LDA, which measu es
he eloci y by means o acking ace s in he suspension. Howe e , PIV measu es he eloci y by
c oss-co ela ion algo i hms, allowing he eloci y 3D-componen s o he eloci y g adien enso in he
measu ing plane o be ob ained [
43
]. PIV has he ad an age o no being a single poin measu emen
echnique, and consequen ly is a ac ing inc easing in e es in he e alua ion o complex and u bulen
lows. The main disad an age o his echnique is he limi a ion i p esen s when mo e opaque domains
a e he objec o isualisa ion. So okin and his co- esea che s used PIV o es ima e se s o 2D pulp
low eloci y ec o s [
8
]. Wi h ecen imp o emen s in digi al imaging, PIV also makes i possible o
ob ain in o ma ion abou he o ien a ion o he ib es, as in Fan e al. [
44
]. In he expe imen s, hey
used syn he ic ib es wi h di e se aspec a ios (18.8, 25.0 and 37.5) in wo di e en concen a ions,
0.1% and 0.2%. They ob ained impo an in o ma ion abou he in luence o he impelle speed, ib e
aspec a io and concen a ion on he ield eloci y and o ien a ion. Fock and colleagues [
45
] used
PIV combined wi h LDA o u he unde s and he mechanisms o plug low and nea -wall beha iou
in ib e low h ough a pipe a concen a ions o up o 4.7% (w/w) wi h a low a e o 0.003 m
3
/s; he
ob ained esul s showed ha he pulp suspension low in he plug egime can be e y inhomogeneous
in he egion nea he pipe wall, pa icula ly a low low eloci ies, becoming mo e s uc u ed wi h an
inc ease in low eloci y.
Among he di icul ield echniques, X- ay compu ed omog aphy is he sa es one; i has excellen
spa ial esolu ion and image econs uc ion is easy, bu due o he sou ce and cons uc ion o de ec ion
componen s, i s ope a ion is slow [
9
]. X- ay usage o ib e low cha ac e isa ion has a ely been
epo ed in he li e a u e; howe e , lash X- ay adiog aphy has been used o cha ac e ise pulp lows
simila o hose ha can be ound in he pape indus y. Flash X- ay is an X- ay adiog aphy me hod
Appl. Sci. 2020,10, 2355 5 o 30
ha uses an in ense bu s o adia ion ha is p oduced o a sho ime pe iod, making i possible o
cap u e images o high-speed e en s ha a e hidden by, o ins ance, dus , smoke o ligh [
9
]. Fa ing on
used Flash X- ay [
46
] in o de o image ungs en ib es wi h a median diame e o 25
µ
m (ac ing as
ace s) in a wood ib e suspension wi h a concen a ion o 2% (w/w). He was able o asses ib e
mass dis ibu ion, con i ming ha his echnique could be used o pulp low cha ac e isa ion. La e ,
Heidel and co-wo ke s [
47
,
48
] used Flash X- ay o cha ac e ise and quan i y gas beha iou in se e al
suspensions con aining ib es; hey used i o isualise ai low in di e se sys ems, pa icula ly in one
composed o ai /wa e /wood ib es, wi h ib e consis encies o 0.5%, 1% and 1.5% (w/w), while a ying
he a e age injec ion eloci y o he ai be ween 0.1 and 12.5 cm/s and main aining he emaining
mix u e injec ion eloci y a a cons an a e in he pipe.
Gamma ay has also been ied o inspec ion o pulp suspensions, pa icula ly o acqui ing
densi y/ oid ac ion in o ma ion [
49
]. Gamma ay densi ome y [
10
] uses Gamma adia ion, which
is an elec omagne ic adia ion simila o X- ays, bu which is p oduced by he na u al decay o
adionuclides. I is eliable o he de e mina ion o densi ies, i is inexpensi e, and i is po able;
howe e , i does no allow local in o ma ion o be ob ained ( he collima ed beam o he Gamma ay
p oduces a line a e aged alue and local in o ma ion in poin s o he cho d canno be ob ained) [
50
].
Also, Gamma ays’ in ensi ies in he domain a e qui e dependen on he hickness o he pipe wall,
and hei accu acy depends on he adia ion eaching he de ec o , which is a ec ed by he ma e ial
and hickness o he pipe wall [
50
]. Xie e al. [
51
] combined Flash X- ay omog aphy wi h Gamma
ay densi ome y o s udy low s uc u es, gas holdup, and he geome y o he gas bubbles in a
gas–pulp–liquid sys em; in hei wo k, low- ib e suspension consis encies we e used—up o 1.5%
(w/w)—and supe icial low eloci ies o gas and liquid/pulp suspensions in he anges 0–26 and 21–51
cm/s, espec i ely.
The s udy o la ge -scale sys ems can be add essed by Elec ical Impedance Tomog aphy (EIT),
which makes i possible o ob ain good images o he dis ibu ion o gas–liquid–solid mix u es [
52
].
Low cos , po abili y and ease o upscaling a e some o he ad an ages o EIT. Howe e , i s
spa ial esolu ion is lowe han ha o o he omog aphic echniques. This is mainly due o he
numbe o ma chless conduc i i y measu emen s ha can be ob ained, which is s ongly ied o he
cu en injec ion- ol age acquisi ion sys em p o ocol and o he numbe o measu ing elec odes [
53
].
The econs uc ion p ocedu e is poo ly condi ioned and non-linea [
53
]. Consequen ly, he amoun
o ib es o pa icles p esen in he suspension canno be p ecisely calcula ed, wi h he ob ained
omog aphic images ep esen ing he dis ibu ion o a ia ion in conduc ance/impedance, a he han
absolu e alues. To maximise he EIT conduc i i y dis ibu ion images esul ing om he econs uc ion
p ocess, he conduc i i y di e ence be ween he wo phases o be dis inguished should be la ge, wi h
la ge conduc i i y di e ences p o iding be e dis inguishabili y and measu emen accu acy. In he
pulp and pape indus y, EIT has ecen ly been used in applica ions including he measu emen o
mixing quali y, low uni o mi y, and eloci y, jus o men ion a ew [11,54–56].
3. Elec ical Tomog aphy P inciples and Applica ions
3.1. Elec ical Tomog aphy
Due o i s low cos , po abili y and handling sa e y (because no ha m ul adia ion is used),
Elec ical Impedance Tomog aphy exhibi s s ong po en ial o use in se e al academic/indus ial
a eas. I has a simple se up, and o e s quick and easy ope a ion, making i obus enough o be
used in many indus ial si ua ions. As has al eady been men ioned, i s main d awback is i s low
spa ial esolu ion, which is usually con ined o be ween 6% and 20% o he pipe adius [
53
]. Howe e ,
wi h he de elopmen o mo e ad anced algo i hms o in e se p oblem esolu ion, he e ha e been
signi ican imp o emen s wi h espec o inc easing he achie able spa ial esolu ion. Wi h Elec ical
Tomog aphy, bo h quan i a i e and quali a i e da a ega ding mul iphase sys ems can be ob ained o
modelling pu poses.

Appl. Sci. 2020,10, 2355 6 o 30
Non-in asi e collec ion o da a by elec ical omog aphy enables he econs uc ion o c oss-sec ion
images, p o iding ma e ial dis ibu ion p o iles in a suspension lowing h ough a pipe, o in o ma ion
abou possible ansien e en s aking place. The esul ing conclusions can hen be used in p ocesses
edesign o con ol. In some cases o he eal- ime imaging o indus ial p ocesses, elec ical omog aphy
is he mos appealing echnique, due o i s simplici y and low cos . Elec ical omog aphy can be
di ided in wo p ocedu es: Elec ical Impedance Tomog aphy, EIT (no e ha Elec ical Resis ance
Tomog aphy, ERT, can be seen has a pa icula case o EIT), and Elec ical Capaci ance Tomog aphy,
ECT. They bo h allow he gene a ion o images ha e lec he a ia ion o he conduc i i y/ esis i i y
o pe mi i i y along he unde -obse a ion domain; due o i s simplici y o implemen a ion, ERT,
which is ideal o pu e esis i e domains, is he mos used [57].
When sys ems a e he e ogeneous, and composed o ma e ials possessing di e se elec ical
cha ac e is ics, EIT and ECT can be used in he obse a ion o changing p ocesses; pneuma ic and
hyd aulic anspo a ion, mixing, cyclones, and luidised beds a e only some examples. The numbe o
images gene a ed by ime in e al is a iable: i depends mainly on he elec onics and s uc u e o he
acquisi ion sys em, on he measu emen p o ocol used o ga he he da a, and on he econs uc ion
algo i hm used o image building. Typically, inc easing he da a ga he ing speed would make i
possible o ob ain da a abou as e p ocess, bu consequen ly wi h a highe noise le el and educed
image quali y o he dis ibu ion p o iles. Elec ical omog aphy can be used on-line o o -line.
Fo on-line measu emen s, poo spa ial esolu ion images a e usually ob ained, consequence o he
need o use ei he as acquisi ion p ocedu es o da a p ocessing algo i hms ha only equi e a sho
ime. Con e sely, o o -line measu emen s, he acquisi ion imes can be cus omised o adap o he
changes aking place in he low p ocess, while he p ocessing algo i hms can be op imised o he
highes spa ial esolu ion possible, because he p ocessing ime can be longe .
In addi ion o he econs uc ion algo i hm used, o he ea u es de e mine he quali y o he
ob ained images, mainly physical ones, such as he numbe o elec odes, elec ode diame e and
ma e ial, spacing be ween hem, e c. Because in some applica ions he aim is o ob ain in o ma ion
abou a e age quan i ies, such as oid ac ion o mean eloci y, he de eloped model’s alida ion
equi es images wi h high spa ial esolu ion.
In he ollowing sec ions, he au ho s s a by add essing he basic p inciples and ope a ion o
Elec ical Impedance Tomog aphy, ollowed by a sho e iew o he use o Elec ical Tomog aphy
o indus ial/p ocess moni o ing pu poses, wi h an emphasis on he pulp and pape a ea. Finally,
a succinc desc ip ion o he au ho s’ app oach o de eloping high spa ial esolu ion po able EIT
equipmen is p esen ed; he au ho s’ EIT sys em makes i possible no only o inspec media wi h
high conduc i i y, as a e ypically ound in indus y, as well as using a econs uc ion p ocedu e ha
makes i possible o ob ain sha pe impedance dis ibu ion images. The esul s ob ained wi h his new
sys em using ib e suspensions on a labo a o y le el pilo ig will be p esen ed. Finally, he incoming
challenges o Elec ical Tomog aphy a e add essed.
3.2. Fundamen als o Elec ical Tomog aphy
Maxwell [
58
] was he i s o p esen a heo e ical model connec ing he pe mi i i y/conduc i i y
o a wo-phase mix u e o he olume ac ion o one he compounds p esen in he mix u e dispe sed
in he second. Fo his calcula ions, he used small sphe es o a ce ain ma e ial possessing equal
bu small size when compa ed wi h he dis ance be ween hem; addi ionally, he p esumed ha
in a ce ain domain, he sphe es we e dis ibu ed in a uni o m way along he con inuous phase
o a second ma e ial, and ha consequen ly hei p esence dis u bs a homogeneous elec ical ield
passing h ough he domain. Ex apola ing, by measu ing quan i ies such as conduc ance/capaci ance,
be ween a pai o elec odes placed a ound he pipe wall, he pa icle dis ibu ion inside ha pipe
can be ob ained by means o he dis ibu ion o he elec ical p ope ies and h ough he usage o an
adequa e ma hema ic p ocedu e: his ma hema ic algo i hm links he measu emen s wi h he pa icle
dis ibu ion. The ela ion be ween he homogenei y o a mul iphase low and his idea has been
Appl. Sci. 2020,10, 2355 7 o 30
used o some ime o obse e di e en sys ems, such as liquid–solid and gas–liquid. Fo ins ance,
o quan i y localised amoun s o solids in a gas low, small capaci i e p obes ha e been used [
59
,
60
].
In he gas and oil indus y, simila p obes ha e been used wi h analogous objec i es [
61
]. Typically,
in elec ical omog aphy, he exci a ion sou ces, ei he a ol age o a cu en sou ce, ope a e a low
equencies (below 5 MHz). Consequen ly, he sys ems in which hese ypes o sou ces a e used
a e desc ibed by he go e ning equa ions o he elec os a ic ield. When he lux (o cu en ) lines
encoun e an in e ace wi h di e en pe mi i i y o conduc i i y, i ge s de lec ed. In o de o allow he
mining o he maximum amoun o in o ma ion abou he p ocess, he elec odes a e usually moun ed
o e he pe iphe y o he pipe a in e als o equal dis ance. In he case o capaci i e sys ems, he
measu ing elec odes a e ins alled ou side o he p ocess essel o pipe which is made o a dielec ic
ma e ial (i.e., hey a e moun ed in a non-in asi e way). Fu he mo e, he elec ode a ea mus be high,
making i possible o de ec a la ge ange o capaci ance changes. Fo sys em measu ing Resis i i y
changes, he sensing elec odes a e qui e small when compa ed o hose used in capaci i e sys ems.
They will ypically be in con ac wi h he low unde inspec ion, moun ed le el along he inne side o
he pipe wall.
3.2.1. Capaci ance Tomog aphy
The pu pose o ECT i o ob ain a econs uc ion o he dielec ic cha ac e is ics o a sys em,
using measu es o capaci ance aken be ween all exis ing elec ode pai s. In Figu e 1, a schema ic o
a c oss-sec ion ep esen a ion o an ECT sys em possessing eigh elec odes can be seen. I he pipe
whe e he suspension is lowing is made o a conduc i e ma e ial, hen he elec odes a e in e nally
moun ed. I , con e sely, he pipe is made o an insula ing ma e ial, hen he elec odes will be loca ed
ex e nally. Ex e nal elec odes a e easily concei ed, assembled and main ained; addi ionally, once
hey ha e been di ec ly exposed o ex eme u bulence, p essu es and empe a u es, hei elec ical
beha iou emains unchanged o a longe ime (con amina ion by he ma e ials lowing in he ube is
no encoun e ed, ei he ). Thei g ea es disad an age is ha hey possess non-linea cha ac e is ics;
howe e , he use o adequa e co ec ion ac o s allows hem o display almos linea cha ac e is ics.
The design o in e nal elec odes is ha de and mo e in ica e, because hey a e exposed as desc ibed
abo e o ex eme condi ions, and may also su e om co osion; howe e , he capaci ance changes
obse ed wi h in e nal elec odes can be assumed o be di ec ly p opo ional o he pe mi i i y
changes occu ing inside he pipe. In ECT, all elec odes a e s imula ed, one a a ime, and all pai s
o capaci ances be ween he s imula ed elec ode and he emaining ones a e measu ed; o an N
elec odes sys em, his p oduces N(N-1)/2 independen measu emen s, once capaci ance C
i,j
=C
j,i
and C
i,i
, i.e., he sel -capaci ance, a e all igno ed. In summa y, he measu ing p o ocol ope a es in a
simila way o he sou ce-de ec o mo emen in compu e ised omog aphy in medical imaging; in
ECT, he elec ical ield o a es a ound he pipe c oss-sec ion in s eps, wi h a s ep angle o 360
◦
/N.
The numbe No elec odes is chosen based on a balance be ween he desi ed spa ial esolu ion and
acquisi ion a e o he images (acquisi ion a es a ound 100 ames pe second a e equen ). Maxwell
equa ions make i possible o es ablish a ela ionship be ween he measu ed capaci ances and he
ac ual spa ial dis ibu ion o he pe mi i i y. In ECT, only one elec ode is s imula ed a one ime,
while he emaining a e a a i ual ea h po en ial. Consequen ly, o e all elec ode su aces, he
o al elec ic lux is equal o ze o. The capaci ance measu es can be easily in luenced by ex e nal
pa asi e capaci ances (usually g ea e han he measu ed ones), such as ea h capaci ance, because
he in e -elec ode capaci ance is qui e small. To p e en his pe u ba ion, elec ode shielding is
implemen ed. Wi h espec o he exci a ion signal equency, a alue o a ound 1 MHz is common; o
his o de o magni ude, he signal wa eleng h is o some hund eds o me e s, g ea ly exceeding he
senso size by many o de s o magni ude. In summa y, elec os a ic ield heo y can be desc ibed by
he elec ical po en ial dis ibu ion inside he medium unde inspec ion.
Appl. Sci. 2020,10, 2355 8 o 30
Appl. Sci. 2020, 10, x FOR PEER REVIEW 7 o 31
conduc i i y, i ge s de lec ed. In o de o allow he mining o he maximum amoun o in o ma ion
abou he p ocess, he elec odes a e usually moun ed o e he pe iphe y o he pipe a in e als o
equal dis ance. In he case o capaci i e sys ems, he measu ing elec odes a e ins alled ou side o he
p ocess essel o pipe which is made o a dielec ic ma e ial (i.e., hey a e moun ed in a non-in asi e
way). Fu he mo e, he elec ode a ea has o be high, making i possible o de ec a la ge ange o
capaci ance changes. Fo sys em measu ing Resis i i y changes, he sensing elec odes a e qui e small
when compa ed o hose used in capaci i e sys ems. They will ypically be in con ac wi h he low
unde inspec ion, moun ed le el along he inne side o he pipe wall.
3.2.1. Capaci ance Tomog aphy
The pu pose o ECT i o ob ain a econs uc ion o he dielec ic cha ac e is ics o a sys em, using
measu es o capaci ance aken be ween all exis ing elec ode pai s. In Figu e 1, a schema ic o a c oss-
sec ion ep esen a ion o an ECT sys em possessing eigh elec odes can be seen. I he pipe whe e he
suspension is lowing is made o a conduc i e ma e ial, hen he elec odes a e in e nally moun ed. I ,
con e sely, he pipe is made o an insula ing ma e ial, hen he elec odes will be loca ed ex e nally.
Ex e nal elec odes a e easily concei ed, assembled and main ained; addi ionally, once hey ha e been
di ec ly exposed o ex eme u bulence, p essu es and empe a u es, hei elec ical beha iou emains
unchanged o a longe ime (con amina ion by he ma e ials lowing in he ube is no encoun e ed,
ei he ). Thei g ea es disad an age is ha hey possess non-linea cha ac e is ics; howe e , he use o
adequa e co ec ion ac o s allows hem o display almos linea cha ac e is ics. The design o in e nal
elec odes is ha de and mo e in ica e, because hey a e exposed as desc ibed abo e o ex eme
condi ions, and may also su e om co osion; howe e , he capaci ance changes obse ed wi h in e nal
elec odes can be assumed o be di ec ly p opo ional o he pe mi i i y changes occu ing inside he
pipe. In ECT, all elec odes a e s imula ed, one a a ime, and all pai s o capaci ances be ween he
s imula ed elec ode and he emaining ones a e measu ed; o an N elec odes sys em, his p oduces N(N-
1)/2 independen measu emen s, once capaci ance Ci,j = Cj,i and Ci,i, i.e., he sel -capaci ance, a e all igno ed.
In summa y, he measu ing p o ocol ope a es in a simila way o he sou ce-de ec o mo emen in
compu e ised omog aphy in medical imaging; in ECT, he elec ical ield o a es a ound he pipe c oss-
sec ion in s eps, wi h a s ep angle o 360°/N. The numbe N o elec odes is chosen based on a balance
be ween he desi ed spa ial esolu ion and acquisi ion a e o he images (acquisi ion a es a ound 100
ames pe second a e equen ). Maxwell equa ions make i possible o es ablish a ela ionship be ween
he measu ed capaci ances and he ac ual spa ial dis ibu ion o he pe mi i i y. In ECT, only one
elec ode is s imula ed a one ime, while he emaining a e a a i ual ea h po en ial. Consequen ly, o e
all elec ode su aces, he o al elec ic lux is equal o ze o. The capaci ance measu es can be easily
in luenced by ex e nal pa asi e capaci ances (usually g ea e han he measu ed ones), such as ea h
capaci ance, because he in e -elec ode capaci ance is qui e small. To p e en his pe u ba ion, elec ode
shielding is implemen ed. Wi h espec o he exci a ion signal equency, a alue o a ound 1 MHz is
common; o his o de o magni ude, he signal wa eleng h is o some hund eds o me e s, g ea ly
exceeding he senso size by many o de s o magni ude. In summa y, elec os a ic ield heo y can be
desc ibed by he elec ical po en ial dis ibu ion inside he medium unde inspec ion.
Figu e 1.
Schema ic ep esen a ion o he measu emen p inciple o an ECT sys em ( ep oduced wi h
pe mission om Hansen e al. [62]).
ERT sensing a ays a e less complex han hose equi ed by ECT, and he di icul y in hei
concep ion a ises when he domains p esen high conduc i i y [
63
,
64
]. Consequen ly, i can be s a ed
ha i is qui e adequa e o he inspec ion o non-conduc i e mix u es con aining di e se pe mi i i y’s.
ECT p oduces low- esolu ion images, bu , when dealing wi h dispe se sys em lows, o e s good
accu acy wi h espec o olume ac ion es ima ion [
65
]. I in ECT a dual plane sys em is used, hen
c oss co ela ing he c oss-sec ion a e age ime se ies enables he e alua ion o a pa icle’s eloci y [
65
].
3.2.2. Resis i i y/Impedance Tomog aphy
The impedance a ia ion dis ibu ion o a mix u e is he ocus o EIT. In EIT, he s imulus used is a
sinusoidal elec ical signal (cu en o ol age) wi h a ce ain equency and ampli ude. Then, he eal
and imagina y pa s o he impedances a e calcula ed using phase sensi i i y measu emen s [
66
,
67
].
The esis i e componen is ob ained by an in-phase measu emen wi h he s imulus signal, while he
capaci i e componen is ound by a quad a u e o phase measu emen wi h he exci a ion signal [
53
].
The in en ion c edi o EIT is gi en o John G. Webs e , who epo ed i in a 1978 publica ion [
68
].
S ill, he i s known expe imen al applica ion o EIT ook place in 1984, o which he au ho s
we e Ba be and B own [
69
]. Fo in es iga ing p ocesses whe e he con inuous phase is elec ically
conduc i e, ei he ERT o EIT can be used. Speci ically, when he inspec ed medium is pu ely esis i e,
hen ERT should be used. Simila ly o ECT, in ERT/EIT an elec ical cu en is injec ed h ough a pai o
elec odes ha a e placed along he bounda y o he domain unde s udy. Consequen ly, he esul ing
elec ical ield dis ibu ion in he domain will be condi ioned by he ma e ial dis ibu ion wi hin i [
70
].
The po en ial ol age di e ences be ween all pai s o elec odes placed a ound he domain
pe ime e , apa om he pai used o exci ing he domain, a e measu ed, and used o nou ish a
non-linea algo i hm. This algo i hm sol es he in e se p oblem desc ibed abo e and pe mi s he
calcula ion o he unknown conduc i i y/ esis i i y dis ibu ion along he domain. Howe e , he
p ocedu e only ends when all elec odes a e used o s imula ing he domain, and hus a ull cycle has
as many p ojec ions as he numbe o elec odes comp ising he sensing sys em (see Figu e 2) [
71
].
The e o e, in summa y, o es ima ing he ma e ial dis ibu ion in he domain, an elec ical ield
dis ibu ion cha ac e isa ion is used. In ERT, which is easie o implemen [
67
,
72
], he aim is o ob ain
a conduc i i y/ esis i i y dis ibu ion along he domain. The wo king p inciple used in ERT o ob ain
he da a ela ed o a ull se o p ojec ions, is iden ical o ha desc ibed o EIT ( equi ing he use
o all elec odes o in oduce he elec ical exci a ion signal in he domain, and o measu e all he
po en ial ol age di e ences be ween he emaining ones). The single di e ence is ha in ERT, he
elec ical signal used as s imulus is a con inuous cu en . The e o e, o EIT, he da a acquisi ion cycle
only ends when a ull o a ion o he applied elec ic ield exci a ion is comple e. Consequen ly, o
bo h EIT and ERT, i he sensing sys em is composed o Nelec odes, he numbe o independen
di e en ial ol age measu emen s is N(N
−
3)/2. Typical alues o he equencies used in EIT sys ems
lie in he ange 20–150 kHz, so quasi-s a ic condi ions can be assumed when a ull o a ion da a is
Appl. Sci. 2020,10, 2355 9 o 30
acqui ed. In ecen yea s, e o s ha e been di ec ed owa ds he de elopmen o acquisi ion sys ems
and o he concep ion o senso s (ma e ials used in hei ab ica ion, dimensions and loca ion o he
senso s, e c.). The use o a disc e ised elec ode sys em is es ic ed o based lows wi h cons an
conduc ance [
73
]. When la ge bubbles a e p esen in he luid, hose bubbles can cause elec odes o
lose con ac wi h he luid, and consequen ly, he ob ained images o he conduc ance dis ibu ion will
be inconsis en . When aced wi h his ype o suspension, one possible solu ion would be he use o
con inuous elec ode ings ins ead o a disc e e elec ode sensing sys em [
73
]. An addi ional limi a ion
o using disc e e elec odes moun ed in he pipe wall, he su ace o which is pa allel o he inne
pipe wall (as equi ed by esis i e/impedance omog aphy), is hei su ace modi ica ion o e ime,
pa icula ly when used o e alua ion p ocesses ha ope a e unde ex eme condi ions. Tempe a u e
is one example o such condi ions: empe a u e changes can d ama ically cause luid conduc i i y o
al e , and consequen ly luid empe a u e compensa ion mechanisms ha e o be applied o “co ec ”
cu en / ol age measu emen s.
Appl. Sci. 2020, 10, x FOR PEER REVIEW 9 o 31
o such condi ions: empe a u e changes can d ama ically cause luid conduc i i y o al e , and
consequen ly luid empe a u e compensa ion mechanisms ha e o be applied o “co ec ”
cu en / ol age measu emen s.
Looking now o he e olu ion o acquisi ion sys ems, ecen de elopmen s in he elec onics ield
ha e made i possible o achie e ame acquisi ion speeds o 1000 ames/s [73,74]. In acquisi ion sys ems,
a good exci a ion s a egy is also i al, and hese a e ypically cu en -based ones. Howe e , he use o
ol age-based exci a ion s a egies, whe e a ol age-con olled sou ce ( ypically wi h low ou pu
impedance) combined wi h an equal wid h pulse syn hesise , gene a es a subse o wa e o ms used o
exci ing he domain and o implemen demodula ion, ha e a ac ed inc easing a en ion, since hey
p oduce less noisy da a [73]. The use o ol age-con olled sou ces in he domain exci a ion causes
elec ical cu en s inside he domain ha inc ease wi h he inc ease in suspension luid conduc i i y; his
ac is e y impo an o highly conduc i e suspensions, abo e 2 S/m, which a e qui e ypical o indus ial
p ocesses [67,74].
Figu e 2. EIT/ERT adjacen injec ion and measu emen p o ocol o he i s (A) and second (B)
p ojec ions ( ep oduced wi h pe mission om Malmi uo e al. [71]).
In addi ion o he shape and size o he elec odes, and he ma e ial om which hey a e made,
when using ERT/EIT, o he cha ac e is ics mus be conside ed, such as he spacing be ween
elec odes, and o EIT he equency o he elec ical exci a ion signal. Al oge he , hese
cha ac e is ics condi ion he elec ic ield dis ibu ion along he domain and de e mine he ue
measu ing olume and o e all sensi i i y [74,75]. Howe e , some o he nega i e e ec s o hese
cha ac e is ics can be o e come; o ins ance, he e ec o he elec ode dimensions can be educed
by using app op ia e calib a ion p ocedu es [76]. EIT, simila o ECT, can also be used o ob ain
dispe se phase eloci y by using a pai o da a acquisi ion planes and applying c oss-co ela ion
algo i hms o he acqui ed da a [77,78]. Howe e , o p ope ly es ima e eloci y, a minimum ame
a e acquisi ion o 100 ames pe second is needed [78]. Demodula ion has o be pe o med h ough
so wa e o o e come he p oblem o he low se ling ime o he low-pass il e s.
The way elec odes a e combined o in oduce he elec ical signal in o he domain is designa ed
by he Injec ion P o ocol. Tho ough s udies on injec ion and measu emen p o ocols o pa e ns can
be ound in he li e a u e. The mos common s a egy is e e ed o as he ou -elec ode me hod,
whe e he signal is injec ed h ough some elec odes and he ol ages a e measu ed om he
emaining elec odes. Di e en pa e ns o using his s a egy can be ound in he li e a u e [79,80]:
(a) Adjacen Pa e n: whe e he signal is injec ed h ough neighbou ing o adjacen elec odes and
he esul ing elec ical po en ial di e ences a e measu ed in he emaining pai s o elec odes, in a
simila adjacen combina ion; (b) Opposi e Pa e n: In his pa e n, he signal is injec ed h ough
opposi e elec odes and he elec ical po en ial is measu ed wi h espec o one e e ence elec ode
adjacen o he injec ing elec ode. In addi ion o he abo e-men ioned s a egies, which a e he mos
equen ly used, o he s ha e been employed, such as he C oss o Diagonal Pa e n [80], which is a
combina ion o he wo desc ibed abo e, and he T igonome ic o Adap i e Pa e n [81].
Figu e 2.
EIT/ERT adjacen injec ion and measu emen p o ocol o he i s (
A
) and second (
B
)
p ojec ions ( ep oduced wi h pe mission om Malmi uo e al. [71]).
Looking now o he e olu ion o acquisi ion sys ems, ecen de elopmen s in he elec onics ield
ha e made i possible o achie e ame acquisi ion speeds o 1000 ames/s [
73
,
74
]. In acquisi ion
sys ems, a good exci a ion s a egy is also i al, and hese a e ypically cu en -based ones. Howe e ,
he use o ol age-based exci a ion s a egies, whe e a ol age-con olled sou ce ( ypically wi h low
ou pu impedance) combined wi h an equal wid h pulse syn hesise , gene a es a subse o wa e o ms
used o exci ing he domain and o implemen demodula ion, ha e a ac ed inc easing a en ion, since
hey p oduce less noisy da a [
73
]. The use o ol age-con olled sou ces in he domain exci a ion causes
elec ical cu en s inside he domain ha inc ease wi h he inc ease in suspension luid conduc i i y;
his ac is e y impo an o highly conduc i e suspensions, abo e 2 S/m, which a e qui e ypical o
indus ial p ocesses [67,74].
In addi ion o he shape and size o he elec odes, and he ma e ial om which hey a e made,
when using ERT/EIT, o he cha ac e is ics mus be conside ed, such as he spacing be ween elec odes,
and o EIT he equency o he elec ical exci a ion signal. Al oge he , hese cha ac e is ics condi ion
he elec ic ield dis ibu ion along he domain and de e mine he ue measu ing olume and o e all
sensi i i y [
74
,
75
]. Howe e , some o he nega i e e ec s o hese cha ac e is ics can be o e come;
o ins ance, he e ec o he elec ode dimensions can be educed by using app op ia e calib a ion
p ocedu es [
76
]. EIT, simila o ECT, can also be used o ob ain dispe se phase eloci y by using a
pai o da a acquisi ion planes and applying c oss-co ela ion algo i hms o he acqui ed da a [77,78].
Howe e , o p ope ly es ima e eloci y, a minimum ame a e acquisi ion o 100 ames pe second is
needed [
78
]. Demodula ion has o be pe o med h ough so wa e o o e come he p oblem o he low
se ling ime o he low-pass il e s.
Appl. Sci. 2020,10, 2355 16 o 30
pulp) and obse ed ha as hei size inc eased o accommoda e highe p oduc ion a es, ope a ing
p oblems ela ed o liquo low inc eased (as de ined by he au ho s, liquo low is he ela i e low o
liquo h ough he chip mass). Among he di e en p oblems, hey ound dec eased coun e cu en
washing e iciency, accele a ed co osion, inc eased pulp non-uni o mi y and p oblems wi h chip
mo emen . In a column a anged o ope a e as he wash sec ion o a con inuous diges e , hey used
ERT o ob ain images o he low p ope ies. Fo he es s ca ied, he packed bed was modelled using
HDPE pelle s o high-densi y polye hylene, ha ing diame e s o a ound 4.44 mm and hickness o
a ound 1.91 mm. Rega ding he liquid phase, a spli be ween liquo lows was made o ob ain a ios
o axial (up low in all cases), Q
A
, o adial low, Q
R
, o 0.5, 1.0 and 2.0. The es s wi h wood chips
showed ha liquid low adjacen o he essel wall was signi ican ly g ea e han a he cen e o he
essel. This di e ence was no as signi ican o he HDPE disks o which he wall-a ec ed zone is
much smalle and below he spa ial esolu ion o he ERT sys em. When a ace was isokine ically
added, ERT econs uc ed images below he down come showed ha he en i e ace lea ed he
down come when Q
R
was wice he alue o Q
A
; ne e heless, o a 1:1 low a io, some ace was
de ec ed abo e he sc een le el. Figu e 4p esen s some images o he measu emen s conduc ed.
Lee e al. [
56
] used Elec ical esis ance omog aphy (ERT) o e alua e he uni o mi y o liquo low
h ough a po ous media in a labo a o y model diges e unde scaled indus ial condi ions: a 1:15
geome ically scaled essel, a essel o pa icle diame e a io o 93:1 o minimise wall e ec s, and close
app oxima ion o liquo supe icial eloci y and po e Reynolds numbe (po e Reynolds numbe is
based on he capilla y model o low h ough packed beds and accoun s o he po osi y,
ε
, o uosi y,
τ
, and speci ic su ace a ea p esen ed o he low). Thei mo i a ion depa ed om he li e a u e,
whe e he ex en o deligni ica ion in ba ch diges e s a ied as a unc ion o chip loca ion in he essel.
This non-uni o mi y may be exace ba ed by se e al ac o s bu is commonly a ibu ed o poo and/o
non-uni o m liquo low h ough he eac o . Local in e s i ial low eloci ies we e measu ed using
pixel-pixel c oss co ela ion echniques. Tes s we e conduc ed o co e he ange o ypical liquo
spli s, Q
T
/Q
B
( he liquo olume ic low a io be ween he op and bo om o he diges e ) ound
in indus ial uni s: 0.5, 1.0, and 2.0. The esul s showed ha c ea ion o uni o m zones was eadily
achie ed, al hough a ce ain low a e was needed in he uppe sec ion o es ablish uni o mi y due o
asymme ic liquo addi ion o he essel. A mixing egion was obse ed a he sc een ele a ion in he
ou low zone om he diges e . The loca ion o his in e ac ion zone mo ed, depending on he ela i e
liquo lux o he op and bo om o he essel. When local low eloci ies we e measu ed using he
omog aphic da a (equal liquo luxes o he op and bo om ci cula ion zones) no low c ossed he
sc een le el. Thus, a s agna ion poin exis ed a he cen e o he essel nea he plane o he ex ac ion
sc eens. Howe e , by o cing liquo o pass h ough he sc een plane (by di ec ing he low om he
op o he bo om o he diges e and bypassing he ci cula ion sc eens, o ice e sa) du ing pa o he
cook, he spa ial dependence o kappa numbe ( ela ed o he deg ee o cooking) should be educed.
Hui and co-wo ke s [
127
] s udied he e ec o non-New onian heology, including he p esence o
yield s ess, cha ac e is ic o pulp ib e in mixing ope a ions, whe e egions o ac i e mo ion a ound he
impelle s can be c ea ed wi h he ca e n size a ec ing he quali y o mixing a ained. A scaled e sion
o a comme cial axial low impelle was used in a s anda d side-en e ing con igu a ion. Measu ed
ca e n diame e s we e compa ed agains model p edic ions a ailable in he li e a u e. Since he s udied
pulp suspensions we e opaque, ca e n size had o be imaged using indi ec echniques o which wo
me hods we e used, being ERT one o hem ( he o he was ul asonic Dopple elocime y, UDV): ERT
was he p e e ed one due o he ma ked dec ease in he ime equi ed o acqui e da a. The shape o
he ca e n was bes app oxima ed as a unca ed igh -ci cula cylinde and, as expec ed, inc easing
he impelle speed inc eased he ca e n olume. Howe e , he de elopmen o he ca e n olume
wi h inc easing impelle speed was no uni o m, which was a ibu ed o he in e ac ion be ween he
ca e n and he essel walls. A model was de eloped o p edic ing ca e n de elopmen in yield s ess
luids which included in e ac ion wi h essel bounda ies. The p oposed model p edic s well he end

Appl. Sci. 2020,10, 2355 17 o 30
in ca e n olume change wi h inc easing impelle speed al hough he absolu e ca e n olume is only
p edic ed o wi hin 13%.
Appl. Sci. 2020, 10, x FOR PEER REVIEW 16 o 31
om he li e a u e, whe e he ex en o deligni ica ion in ba ch diges e s a ied as a unc ion o chip
loca ion in he essel. This non-uni o mi y may be exace ba ed by se e al ac o s bu is commonly
a ibu ed o poo and/o non-uni o m liquo low h ough he eac o . Local in e s i ial low
eloci ies we e measu ed using pixel-pixel c oss co ela ion echniques. Tes s we e conduc ed o
co e he ange o ypical liquo spli s, QT/QB ( he liquo olume ic low a io be ween he op and
bo om o he diges e ) ound in indus ial uni s: 0.5, 1.0, and 2.0. The esul s showed ha c ea ion o
uni o m zones was eadily achie ed, al hough a ce ain low a e was needed in he uppe sec ion o
es ablish uni o mi y due o asymme ic liquo addi ion o he essel. A mixing egion was obse ed
a he sc een ele a ion in he ou low zone om he diges e . The loca ion o his in e ac ion zone
mo ed, depending on he ela i e liquo lux o he op and bo om o he essel. When local low
eloci ies we e measu ed using he omog aphic da a (equal liquo luxes o he op and bo om
ci cula ion zones) no low c ossed he sc een le el. Thus, a s agna ion poin exis ed a he cen e o
he essel nea he plane o he ex ac ion sc eens. Howe e , by o cing liquo o pass h ough he
sc een plane (by di ec ing he low om he op o he bo om o he diges e and bypassing he
ci cula ion sc eens, o ice e sa) du ing pa o he cook, he spa ial dependence o kappa numbe
( ela ed o he deg ee o cooking) should be educed.
Figu e 4. Tomog aphic images o a liquid wi h a ce ain concen a ion, a liquo , lowing h ough a
packed bed o HDPE pelle s a s eady low condi ions, ob ained using a se o sensing elec ode ings
loca ed sequen ially along a pipe ( he planes a e numbe ed om he bo om, s a ing wi h numbe 1).
A s ep educ ion in liquid conduc i i y was made in he low exi ing he down come . Images display
he liquid concen a ion dis ibu ion a s eady-s a e low a ios o QA/QR o 0.5, 1.0 and 2.0. A o al
liquid lux o 75 mL/s was used o QA/QR o 0.5 and 2.0, and o 50 mL/s o QA/QR equal o 1.0. In he
omog aphic diag ams, low conduc i i y egions a e colou ed blue and high conduc i i y egions a e
colou ed ed (Rep oduced wi h pe mission om Ruzinsky e al. [126]).
Hui and co-wo ke s [127] s udied he e ec o non-New onian heology, including he p esence
o yield s ess, cha ac e is ic o pulp ib e in mixing ope a ions, whe e egions o ac i e mo ion
a ound he impelle s can be c ea ed wi h he ca e n size a ec ing he quali y o mixing a ained. A
scaled e sion o a comme cial axial low impelle was used in a s anda d side-en e ing
con igu a ion. Measu ed ca e n diame e s we e compa ed agains model p edic ions a ailable in he
li e a u e. Since he s udied pulp suspensions we e opaque, ca e n size had o be imaged using
Figu e 4.
Tomog aphic images o a liquid wi h a ce ain concen a ion, a liquo , lowing h ough a
packed bed o HDPE pelle s a s eady low condi ions, ob ained using a se o sensing elec ode ings
loca ed sequen ially along a pipe ( he planes a e numbe ed om he bo om, s a ing wi h numbe 1).
A s ep educ ion in liquid conduc i i y was made in he low exi ing he down come . Images display
he liquid concen a ion dis ibu ion a s eady-s a e low a ios o Q
A
/Q
R
o 0.5, 1.0 and 2.0. A o al
liquid lux o 75 mL/s was used o Q
A
/Q
R
o 0.5 and 2.0, and o 50 mL/s o Q
A
/Q
R
equal o 1.0. In he
omog aphic diag ams, low conduc i i y egions a e colou ed blue and high conduc i i y egions a e
colou ed ed (Rep oduced wi h pe mission om Ruzinsky e al. [126]).
4.3. Pulp Suspensions Cha ac e isa ion Using ET: The Au ho s’ App oach
4.3.1. O e iew o he De eloped EIT Sys em and o he Expe imen al Se up
The au ho s o his pape ha e also been wo king on he imp o emen o a p op ie a y EIT sys em
used in da a acquisi ion o wo-phase lows (solid/liquid). The econs uc ed images aim o alida e
he CFD de eloped model [
74
,
112
,
128
], pa icula ly when applied o he low cha ac e isa ion o pulp
suspensions. To ob ain imp o ed sensi i i y, he de eloped EIT sys em de e mines he di e ences
in bo h he eal and imagina y pa s o he impedance o e he domain. I is a po able de ice;
consequen ly, i can be used in any loca ion whe e 16 o 32 elec ode planes ha e been ins alled on
pipes. To a oid he cos and di icul ies inhe en in he cons uc ion o a cu en sou ce, a ol age sou ce
was p ojec ed ins ead, a choice di e en om wha is ypical when designing elec ical omog aphy
appa a us. Jia e al. [
67
], demons a ed ha he use o a po en ial ol age di e ence ins ead o cu en in
he domain exci a ion p ocedu e made i possible o su pass he limi a ions ound when s udying mo e
conduc i e media. The de eloped EIT sys em has al eady been desc ibed p e iously [
74
]; ne e heless,
oge he wi h a block diag am o he sys em a chi ec u e and i s main modules, p esen ed in Figu e 5, a
b ie desc ip ion is gi en he e. A se o inpu /ou pu boa ds wi h USB po s (DAQ boa ds) a e used o
in e acing he EIT sys em wi h a pe sonal compu e (PC). Using he analogue inpu channels o he
DAQ boa ds, he elec ical po en ials, esul ing om each applied s imulus, a e ead. The de eloped
Appl. Sci. 2020,10, 2355 18 o 30
EIT sys em can acqui e da a a a a e ha allows an image econs uc ion a e o up o 4000 o 1000
ames pe second, depending on he numbe o elec odes composing he sensing ings, ei he 16
o 32. To ob ain he desi ed exci a ion signals, wi h he desi ed ou pu ampli ude and equency (up
o 25 MHz o his sys em), a p og ammable equency ol age wa e o m was used. The au ho s
also adop ed an exci a ion s a egy based on a ol age signal ins ead o a cu en , as has al eady
been men ioned, as he use o his app oach esul ed in less noisy da a [
73
]. In pa allel wi h he
gene a ion o he s imulus signal, wo o he s we e also c ea ed, bo h wi h uni a y magni ude, bu
whe e one is in phase wi h he exci a ion signal and he o he is 90 deg ees ou o phase. To demodula e
he ead elec ical po en ials, each o hem is mul iplied simul aneously by bo h in phase and ou
o phase uni a y p oduced signals; he signals ou pu ed by he mul iplie s go h ough low pass
il e s, enabling only he cons an e ms o be ex ac ed om hem, which a e he e ms needed o
calcula e he impedance dis ibu ion in he domains. A se o analogue bidi ec ional mul iplexe s a e
used o ou ing o all signals: he exci a ion is di ec ed o he desi ed elec ode pai , and he ead
elec ic po en ials o he mul iplie s. The digi al ou pu s o he DAQ boa ds a e used o add ess he
mul iplexe s. This EIT sys em was used o cha ac e ise pulp suspensions in low es s conduc ed in
a pilo ig desc ibed elsewhe e [
129
] and schema ically shown in Figu e 6. In sho , he es sec ion
consis ed o a ho izon al PVC pipe, 7.5 m in leng h and 100 mm in e nal diame e .
Appl. Sci. 2020, 10, x FOR PEER REVIEW 18 o 31
Figu e 5. Block diag am o he ac ual omog aphic sys em main modules.
Figu e 6. Pilo ig schema ic.
The low a e could be egula ed h ough he manipula ion o wo al es loca ed a e he pump
and was measu ed by an elec omagne ic low me e . App op ia e leng hs we e inse ed be o e and
a e he es sec ion o accoun o en ance and exi e ec s. The EIT elec odes ings (cons uc ed in
Te lon), see Figu e 7, consis ed o 32 Ti anium-based elec odes (each wi h a diame e o 5 mm which
was p e iously op imised expe imen ally), ci cum e en ially and equally spaced, and moun ed in a
Pe spex ube inse ed in he ig be ween he measu ing p essu e aps. Fo all es s epo ed he e, an
exci a ion equency o 10 kHz wi h 2Vpp ampli ude was imposed, and opposi e injec ion and adjacen
measu ing p o ocols we e used. Fo image econs uc ion, he open-sou ce so wa e EIDORS [130],
conside ing di ec di e en ial measu ed ol ages and using a s uc u ed Mesh consis ing o 2304
linea elemen s and 1201 nodes, was used. EIDORS implemen s a non-linea back p ojec ion me hod
using a egula ised algo i hm (Tikhono egula isa ion). To sol e he o wa d p oblem he Comple e
Elec ode model (CEM) was chosen [131]. In he econs uc ed images o he conduc i i y dis ibu ion
in he pipe c oss-sec ion, elec ode numbe 1 co esponds o he op posi ion in he c oss-sec ion o
he ube; a da ke blue colou co esponds o a lowe conduc i i y egion, whe e he concen a ion
o he ib es is highe , while yellow o ed colou s iden i y highe conduc i i ies egions, whe e he
ib es a e p esen in lowe concen a ions.
Figu e 5. Block diag am o he ac ual omog aphic sys em main modules.
Appl. Sci. 2020, 10, x FOR PEER REVIEW 18 o 31
Figu e 5. Block diag am o he ac ual omog aphic sys em main modules.
Figu e 6. Pilo ig schema ic.
The low a e could be egula ed h ough he manipula ion o wo al es loca ed a e he pump
and was measu ed by an elec omagne ic low me e . App op ia e leng hs we e inse ed be o e and
a e he es sec ion o accoun o en ance and exi e ec s. The EIT elec odes ings (cons uc ed in
Te lon), see Figu e 7, consis ed o 32 Ti anium-based elec odes (each wi h a diame e o 5 mm which
was p e iously op imised expe imen ally), ci cum e en ially and equally spaced, and moun ed in a
Pe spex ube inse ed in he ig be ween he measu ing p essu e aps. Fo all es s epo ed he e, an
exci a ion equency o 10 kHz wi h 2Vpp ampli ude was imposed, and opposi e injec ion and adjacen
measu ing p o ocols we e used. Fo image econs uc ion, he open-sou ce so wa e EIDORS [130],
conside ing di ec di e en ial measu ed ol ages and using a s uc u ed Mesh consis ing o 2304
linea elemen s and 1201 nodes, was used. EIDORS implemen s a non-linea back p ojec ion me hod
using a egula ised algo i hm (Tikhono egula isa ion). To sol e he o wa d p oblem he Comple e
Elec ode model (CEM) was chosen [131]. In he econs uc ed images o he conduc i i y dis ibu ion
in he pipe c oss-sec ion, elec ode numbe 1 co esponds o he op posi ion in he c oss-sec ion o
he ube; a da ke blue colou co esponds o a lowe conduc i i y egion, whe e he concen a ion
o he ib es is highe , while yellow o ed colou s iden i y highe conduc i i ies egions, whe e he
ib es a e p esen in lowe concen a ions.
Figu e 6. Pilo ig schema ic.
The low a e could be egula ed h ough he manipula ion o wo al es loca ed a e he pump
and was measu ed by an elec omagne ic low me e . App op ia e leng hs we e inse ed be o e and
Appl. Sci. 2020,10, 2355 19 o 30
a e he es sec ion o accoun o en ance and exi e ec s. The EIT elec odes ings (cons uc ed in
Te lon), see Figu e 7, consis ed o 32 Ti anium-based elec odes (each wi h a diame e o 5 mm which
was p e iously op imised expe imen ally), ci cum e en ially and equally spaced, and moun ed in a
Pe spex ube inse ed in he ig be ween he measu ing p essu e aps. Fo all es s epo ed he e, an
exci a ion equency o 10 kHz wi h 2V
pp
ampli ude was imposed, and opposi e injec ion and adjacen
measu ing p o ocols we e used. Fo image econs uc ion, he open-sou ce so wa e EIDORS [
130
],
conside ing di ec di e en ial measu ed ol ages and using a s uc u ed Mesh consis ing o 2304 linea
elemen s and 1201 nodes, was used. EIDORS implemen s a non-linea back p ojec ion me hod using a
egula ised algo i hm (Tikhono egula isa ion). To sol e he o wa d p oblem he Comple e Elec ode
model (CEM) was chosen [
131
]. In he econs uc ed images o he conduc i i y dis ibu ion in he
pipe c oss-sec ion, elec ode numbe 1 co esponds o he op posi ion in he c oss-sec ion o he ube; a
da ke blue colou co esponds o a lowe conduc i i y egion, whe e he concen a ion o he ib es is
highe , while yellow o ed colou s iden i y highe conduc i i ies egions, whe e he ib es a e p esen
in lowe concen a ions.
Appl. Sci. 2020, 10, x FOR PEER REVIEW 19 o 31
Figu e 7. Image o one o he 32 Ti anium elec odes es ings.
4.3.2. O e iew o he Pulp Suspension Flow Imaging
Fo he es s depic ed in his sec ion, indus ial pulp suspensions o Pine and Eucalyp us ib es,
bo h collec ed in a local pape mill, we e i s ly squeezed and washed in wa e , and he ib es we e
hen esuspended in ap wa e doped wi h NaCl wi h adjus ed elec ical conduc i i y.
The beha iou s o Eucalyp us pulp suspensions (a e age Fib e leng h o 0.706 mm) and o Pine
pulp suspensions (a e age Fib e leng h o 2.18 mm) we e compa ed, as in Figu e 8, unde he
condi ions lis ed in Table 1.
Table 1. Table o concen a ions and low eloci ies es ed.
Type o Pulp
Concen a ion (w /w )
Flow Veloci ies (m/s)
Suspension Conduc i i y (mS·cm−1)
Eucalyp
2.35
0.5, 1, 1.5, 2, 2.5, 3
1.460
1.5
0.5, 1, 1.5, 2, 2.5, 3
1.150
1
0.5, 1, 1.5, 2, 2.5, 3
1.128
Pine
2.35
0.5, 1, 1.5, 2, 2.5, 3
1.471
1.5
0.5, 1, 1.5, 2, 2.5, 3
1.461
1
0.5, 1, 1.5, 2, 2.5, 3
1.435
Figu e 7. Image o one o he 32 Ti anium elec odes es ings.
4.3.2. O e iew o he Pulp Suspension Flow Imaging
Fo he es s depic ed in his sec ion, indus ial pulp suspensions o Pine and Eucalyp us ib es,
bo h collec ed in a local pape mill, we e i s ly squeezed and washed in wa e , and he ib es we e
hen esuspended in ap wa e doped wi h NaCl wi h adjus ed elec ical conduc i i y.
The beha iou s o Eucalyp us pulp suspensions (a e age Fib e leng h o 0.706 mm) and o Pine
pulp suspensions (a e age Fib e leng h o 2.18 mm) we e compa ed, as in Figu e 8, unde he condi ions
lis ed in Table 1.
Table 1. Table o concen a ions and low eloci ies es ed.
Type o Pulp Concen a ion
(w /w ) Flow Veloci ies (m/s) Suspension
Conduc i i y (mS·cm−1)
Eucalyp 2.35 0.5, 1, 1.5, 2, 2.5, 3 1.460
1.5 0.5, 1, 1.5, 2, 2.5, 3 1.150
1 0.5, 1, 1.5, 2, 2.5, 3 1.128
Pine 2.35 0.5, 1, 1.5, 2, 2.5, 3 1.471
1.5 0.5, 1, 1.5, 2, 2.5, 3 1.461
1 0.5, 1, 1.5, 2, 2.5, 3 1.435
Be o e analysing he ob ained suspensions low images, we mus ein o ce ha he omog aphy’s
e lec he ol age di e ences measu ed. Fo ib e concen a ions o a ound 2.35%, he ib e dis ibu ions
in he c oss-pipe sec ion a e di e se o he wo ypes o ib es. Conce ning he Pine suspension, a plug
ype low beha iou is al eady isible a he lowes eloci y, wi h an inc ease in he plug (size and
Appl. Sci. 2020,10, 2355 20 o 30
in ensi y o he blue colou ) when he eloci y inc eases o 2 m/s (see Figu e 8a,c,e). Howe e , a u he
inc ease in he eloci y up o 3 m/s leads o b eakage o he plug, as in Figu e 8g), as was expec ed o
his ype o long and s i ib e.
Appl. Sci. 2020, 10, x FOR PEER REVIEW 20 o 31
Figu e 8. EIT images ob ained wi h a concen a ion o ≅2.35% (w/w) o Pine (le column: (a,c,e,g)) and
Eucalyp us ( igh column: (b,d, ,h)) ib es wi h inc easing low eloci y (0.5 m/s, 1m/s, 2m/s and 3m/s,
op o bo om).
Be o e analysing he ob ained suspensions low images, we mus ein o ce ha he
omog aphy’s e lec he ol age di e ences measu ed. Fo ib e concen a ions o a ound 2.35%, he
ib e dis ibu ions in he c oss-pipe sec ion a e di e se o he wo ypes o ib es. Conce ning he
Pine suspension, a plug ype low beha iou is al eady isible a he lowes eloci y, wi h an inc ease
in he plug (size and in ensi y o he blue colou ) when he eloci y inc eases o 2 m/s (see Figu e
8a,c,e). Howe e , a u he inc ease in he eloci y up o 3m/s leads o b eakage o he plug, as in
Figu e 8g), as was expec ed o his ype o long and s i ib e.
When analysing he Eucalyp us ib e suspension beha iou , a he lowes eloci y he ib es
emain mo e o less dispe sed in he c oss-pipe sec ion (see Figu e 8b), and he plug only s a s o be
isible a he eloci y o 1 m/s, becoming dense a 2 m/s (Figu e 8d, ).
Simila ly o he Pine ib es, a u he inc ease in he Eucalyp us suspension eloci y up o 3 m/s
leads o plug b eakage (Figu e 8h). These di e ences in he econs uc ed EIT images o his
concen a ion a e ela ed o he physical p ope ies o he ib es, in pa icula he ib e leng h and
s i ness o he wo pulps ha o igina e di e en in e ac ions be ween ib es when he suspensions
low in he ube.
Fo a ib e concen a ion o a ound 1% (EIT images depic ed in Figu e 9), i was no possible o collec
consis en da a o he Eucalyp us ib es, and so Figu e 9 depic s EIT images only o Pine ib e suspensions.
Fo hese suspensions, and by compa ison wi h Figu e 8, i is e iden ha a lowe concen a ions and
1
2
3
b)
d)
)
g)
h)
e)
c)
a)
Figu e 8.
EIT images ob ained wi h a concen a ion o

2.35% (w/w) o Pine (le column: (
a
,
c
,
e
,
g
)) and
Eucalyp us ( igh column: (
b
,
d
,
,
h
)) ib es wi h inc easing low eloci y (0.5 m/s, 1m/s, 2m/s and 3m/s,
op o bo om).
When analysing he Eucalyp us ib e suspension beha iou , a he lowes eloci y he ib es emain
mo e o less dispe sed in he c oss-pipe sec ion (see Figu e 8b), and he plug only s a s o be isible a
he eloci y o 1 m/s, becoming dense a 2 m/s (Figu e 8d, ).
Simila ly o he Pine ib es, a u he inc ease in he Eucalyp us suspension eloci y up o 3 m/s leads
o plug b eakage (Figu e 8h). These di e ences in he econs uc ed EIT images o his concen a ion
a e ela ed o he physical p ope ies o he ib es, in pa icula he ib e leng h and s i ness o he wo
pulps ha o igina e di e en in e ac ions be ween ib es when he suspensions low in he ube.
Fo a ib e concen a ion o a ound 1% (EIT images depic ed in Figu e 9), i was no possible
o collec consis en da a o he Eucalyp us ib es, and so Figu e 9depic s EIT images only o Pine
ib e suspensions. Fo hese suspensions, and by compa ison wi h Figu e 8, i is e iden ha a lowe
concen a ions and lowe eloci ies (in his case 0.5 m/s), Pine ib es mig a e and become agglome a ed
in a cen al plug, while o a mo e a mo e concen a ed suspension, he ib es (Figu e 8) occupy
a la ge egion o he pipe c oss-sec ion. Howe e , a mo e dilu ed suspensions, he s eng h o
Appl. Sci. 2020,10, 2355 21 o 30
he in e connec ion o he ib es is lowe and he plug b eakage occu s a lowe eloci ies as he
eloci y inc eases.
Appl. Sci. 2020, 10, x FOR PEER REVIEW 21 o 31
lowe eloci ies (in his case 0.5 m/s), Pine ib es mig a e and become agglome a ed in a cen al plug, while
o a mo e a mo e concen a ed suspension, he ib es (Figu e 8) occupy a la ge egion o he pipe c oss-
sec ion. Howe e , a mo e dilu ed suspensions, he s eng h o he in e connec ion o he ib es is lowe
and he plug b eakage occu s a lowe eloci ies as he eloci y inc eases.
Figu e 9. EIT images ob ained wi h a concen a ion o ≅1% (w/w) o Pine ib es (a–d) o 0.5 m/s, 1
m/s, 1.5 m/s and 2.5 m/s, espec i ely.
I mus be emphasised a his poin ha en i onmen al ac o s ha we e no easily kep unde
con ol may a ec he colou mapping o he econs uc ed images. In pa icula , he empe a u e
changes signi ican ly du ing con inuous es ing due o he ene gy inpu om he pump. Mo eo e ,
since cooling o he sys em was pe o med by passing ap wa e h ough a ube imme sed in he
suspension ank, he empe a u e o his cooling wa e changes signi ican ly wi h espec o he
seasons, which also a ec s he empe a u e in he ig. Changes in he suspension empe a u e a ec
he iscosi y, bu mainly he elec ic conduc i i y o he suspension and, consequen ly, he colou
map o he econs uc ed images is a ec ed.
4.3.3. Fib e Flow Veloci y Es ima ion
By using wo EIT planes and c oss-co ela ion, EIT can also be used o de e mine he eloci y
p o ile in a pipe. Fo example, Moso o e al. [132] used a bes -co ela ed pixel me hod o measu e
bo h axial and adial eloci ies simul aneously. Independen ly o he p ocedu e used, a ma ke is
needed. As a i s app oach, he in oduc ion o a small ac ion o conduc i e o non-conduc i e solid
sphe es in he low was used. A e se e al a emp s wi hou success o dis inguish hese ma ke s in
he EIT images, he au ho s ins ead decided o use a ace o a highly concen a ed NaCl solu ion as
a ma ke . Fo he pu pose o in oducing his NaCl solu ion in he lowing ib e suspension, an
injec ion sys em was de eloped and added o he exis ing pilo ig: i was loca ed a ound 3 m be o e
he i s EIT elec ode ing. This injec o was buil in such a way ha he low su e ed almos no
dis u bance. Fo all he es s epo ed in his sec ion, wo 16 elec ode i anium ings A and B we e
used, 30 cm apa , wi h ing A close o he injec o inle , as shown in Figu e 10. The p ocedu e he e
epo ed o es ima ing eloci y was di e en om hose epo ed in he li e a u e, bu , ne e heless,
wi h some simila i ies wi h he one desc ibed by Dong e al. [133]. Ins ead o using he comple e se
o EIT p ojec ions, he au ho s only made use o he i s p ojec ion; i s , he po en ial di e ences o
each pai o he i s p ojec ion wi h no ma ke injec ion in he low we e acqui ed, a e wa ds, hey
we e sub ac ed o he measu ed ol age di e ences ob ained du ing he uns wi h NaCl injec ion.
1
2
d)
b)
a)
c)
Figu e 9.
EIT images ob ained wi h a concen a ion o

1% (w/w) o Pine ib es (
a
–
d
) o 0.5 m/s, 1 m/s,
1.5 m/s and 2.5 m/s, espec i ely.
I mus be emphasised a his poin ha en i onmen al ac o s ha we e no easily kep unde
con ol may a ec he colou mapping o he econs uc ed images. In pa icula , he empe a u e
changes signi ican ly du ing con inuous es ing due o he ene gy inpu om he pump. Mo eo e ,
since cooling o he sys em was pe o med by passing ap wa e h ough a ube imme sed in he
suspension ank, he empe a u e o his cooling wa e changes signi ican ly wi h espec o he
seasons, which also a ec s he empe a u e in he ig. Changes in he suspension empe a u e a ec
he iscosi y, bu mainly he elec ic conduc i i y o he suspension and, consequen ly, he colou map
o he econs uc ed images is a ec ed.
4.3.3. Fib e Flow Veloci y Es ima ion
By using wo EIT planes and c oss-co ela ion, EIT can also be used o de e mine he eloci y
p o ile in a pipe. Fo example, Moso o e al. [
132
] used a bes -co ela ed pixel me hod o measu e
bo h axial and adial eloci ies simul aneously. Independen ly o he p ocedu e used, a ma ke is
needed. As a i s app oach, he in oduc ion o a small ac ion o conduc i e o non-conduc i e solid
sphe es in he low was used. A e se e al a emp s wi hou success o dis inguish hese ma ke s
in he EIT images, he au ho s ins ead decided o use a ace o a highly concen a ed NaCl solu ion
as a ma ke . Fo he pu pose o in oducing his NaCl solu ion in he lowing ib e suspension, an
injec ion sys em was de eloped and added o he exis ing pilo ig: i was loca ed a ound 3 m be o e
he i s EIT elec ode ing. This injec o was buil in such a way ha he low su e ed almos no
dis u bance. Fo all he es s epo ed in his sec ion, wo 16 elec ode i anium ings A and B we e
used, 30 cm apa , wi h ing A close o he injec o inle , as shown in Figu e 10. The p ocedu e he e
epo ed o es ima ing eloci y was di e en om hose epo ed in he li e a u e, bu , ne e heless,
wi h some simila i ies wi h he one desc ibed by Dong e al. [
133
]. Ins ead o using he comple e se o
EIT p ojec ions, he au ho s only made use o he i s p ojec ion; i s , he po en ial di e ences o each
pai o he i s p ojec ion wi h no ma ke injec ion in he low we e acqui ed, a e wa ds, hey we e
sub ac ed o he measu ed ol age di e ences ob ained du ing he uns wi h NaCl injec ion.
Finally, he sum o he module o hose di e ences was compu ed. Fo each suspension eloci y
and ib e concen a ion es ed, se e al injec ion ials we e pe o med o assess he ep oducibili y o his
app oach. The condi ions used o he eloci y es ima ion a e summa ised in Table 2. Some examples o
da a ob ained wi h his app oach a e shown in Figu e 11: h ee se s o NaCl injec ions o suspensions
con aining 1% and 2% o Eucalyp us ib es wi h a eloci y o 0.25 m/s (Figu e 11a,b).

Appl. Sci. 2020,10, 2355 22 o 30
Appl. Sci. 2020, 10, x FOR PEER REVIEW 22 o 31
Figu e 10. Schema ic o he ela i e posi ion o he EIT elec odes sec ions and NaCl injec ion inle .
Finally, he sum o he module o hose di e ences was compu ed. Fo each suspension eloci y
and ib e concen a ion es ed, se e al injec ion ials we e pe o med o assess he ep oducibili y o
his app oach. The condi ions used o he eloci y es ima ion a e summa ised in Table 2. Some
examples o da a ob ained wi h his app oach a e shown in Figu e 11: h ee se s o NaCl injec ions
o suspensions con aining 1% and 2% o Eucalyp us ib es wi h a eloci y o 0.25 m/s (Figu e 11a,b).
Table 2. Table o ib e concen a ions and low eloci ies used o he eloci y es ima ion es s.
Concen a ion (w /w )
Flow Veloci ies (m/s)
3
0.25, 0.5
2
0.25, 0.5, 0.75, 1
1
0.25, 0.5, 0.75, 1
In ha same igu e, a g aph showing he es ima ed eloci y e sus he measu ed suspension
eloci y is ep esen ed (Figu e 11c). A qui e easonable es ima ion is ob ained, and so he au ho s’
app oach was alida ed. Mo eo e , as ib e concen a ion inc eases, he ad ance o he NaCl on is
slowed due o he p esence o mo e ib es, and also due o he s uc u es ha build up among ib es.
This can be obse ed be ween Figu es 11a and b, whe e o he highe ib e concen a ion, he e is a
delay in he de ec ion o he NaCl in he second ing ( ing B). The e o e, in he u u e, i can be used
o e alua e hose s uc u es as a unc ion o ib e concen a ion and low eloci y.
Figu e 10. Schema ic o he ela i e posi ion o he EIT elec odes sec ions and NaCl injec ion inle .
Table 2. Table o ib e concen a ions and low eloci ies used o he eloci y es ima ion es s.
Concen a ion (w /w ) Flow Veloci ies (m/s)
3 0.25, 0.5
2 0.25, 0.5, 0.75, 1
1 0.25, 0.5, 0.75, 1
Appl. Sci. 2020, 10, x FOR PEER REVIEW 23 o 31
Figu e 11. EIT injec ion es ials o ib e suspension concen a ions o 1% and 2%, a he eloci y o
0.25 m/s (a,b), and es ima ed eloci y e sus measu ed eloci y o all ib e concen a ions (c).
5. Conclusions
Theo e ical models de eloped o desc ibe mul iphase lows ha e been de eloped ha ha e
di e en assump ions equi ing expe imen al e i ica ion o alida ion. Fo example, eliable da a
on eloci y p o iles in concen a ed pulp suspensions du ing pipe low is sca ce in he li e a u e. A
mul i ude o echniques is now a ailable, bu a condi ion o non-dis u bance o he lows is equi ed
in iew o he complexi y o mul iphase sys ems. Non-in asi eness is he e o e bene icial, and is
cha ac e is ic o some o hese echniques. Lase Dopple Anemome y (LDA), Magne ic Resonance
Imaging (MRI), Ul asonic Veloci y P o iling (UVP) and Elec ical Tomog aphy we e e iewed and
compa ed o hei ene gy equi emen s, spa ial and empo al esolu ion, du a ion o obse a ion
imes, spa ial and empo al cohe ence and s abili y, cos o equipmen , and ield and scale o
applica ions. The comp omise be ween he ad an ages and limi a ions o hese echniques led us o
ocus on he Elec ic Tomog aphic echniques, he p inciples behind hem, and hei applica ion in
he isualisa ion o pulp suspension lows. Ou e o s we e di ec ed owa ds designing p op ie a y,
low-cos , po able Elec ic Impedance Tomog aphic (EIT) ha dwa e. To ci cum en he di icul ies
and cos s o designing a cu en sou ce ha sui s he goals o he EIT sys em being designed, he
choice was made o mo e om adi ional EIT sys ems and, ins ead, o design a ol age sou ce. The
EIT sys em depic ed he e can be used wi h 16-elec ode o 32-elec ode ings. Examples o Eucalyp us
and Pine pulp suspension lows a e epo ed he e, demons a ing he in luence o ib e concen a ion
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Injec ion 1
Injec ion 2
Injec ion 1
Injec ion 2
Ring A de ec ion
Ring B de ec ion
a)
b)
c)
Figu e 11.
EIT injec ion es ials o ib e suspension concen a ions o 1% and 2%, a he eloci y o
0.25 m/s (a,b), and es ima ed eloci y e sus measu ed eloci y o all ib e concen a ions (c).
In ha same igu e, a g aph showing he es ima ed eloci y e sus he measu ed suspension
eloci y is ep esen ed (Figu e 11c). A qui e easonable es ima ion is ob ained, and so he au ho s’
app oach was alida ed. Mo eo e , as ib e concen a ion inc eases, he ad ance o he NaCl on is
slowed due o he p esence o mo e ib es and o he s uc u es ha build up among ib es. This can
Appl. Sci. 2020,10, 2355 23 o 30
be obse ed be ween Figu e 11a,b, whe e o he highe ib e concen a ion, he e is a delay in he
de ec ion o he NaCl in he second ing ( ing B). The e o e, in he u u e, i can be used o e alua e
hose s uc u es as a unc ion o ib e concen a ion and low eloci y.
5. Conclusions
Theo e ical models de eloped o desc ibe mul iphase lows ha e been de eloped ha ha e
di e en assump ions equi ing expe imen al e i ica ion o alida ion. Fo example, eliable da a
on eloci y p o iles in concen a ed pulp suspensions du ing pipe low is sca ce in he li e a u e.
A mul i ude o echniques is now a ailable, bu a condi ion o non-dis u bance o he lows is equi ed
in iew o he complexi y o mul iphase sys ems. Non-in asi eness is he e o e bene icial and is
cha ac e is ic o some o hese echniques. Lase Dopple Anemome y (LDA), Magne ic Resonance
Imaging (MRI), Ul asonic Veloci y P o iling (UVP) and Elec ical Tomog aphy we e e iewed and
compa ed o hei ene gy equi emen s, spa ial and empo al esolu ion, du a ion o obse a ion imes,
spa ial and empo al cohe ence and s abili y, cos o equipmen , and ield and scale o applica ions.
The comp omise be ween he ad an ages and limi a ions o hese echniques led us o ocus on he
Elec ic Tomog aphic echniques, he p inciples behind hem, and hei applica ion in he isualisa ion
o pulp suspension lows. Ou e o s we e di ec ed owa ds designing p op ie a y, low-cos , po able
Elec ic Impedance Tomog aphic (EIT) ha dwa e. To ci cum en he di icul ies and cos s o designing
a cu en sou ce ha sui s he goals o he EIT sys em being designed, he choice was made o mo e
om adi ional EIT sys ems and, ins ead, o design a ol age sou ce. The EIT sys em depic ed he e
can be used wi h 16-elec ode o 32-elec ode ings. Examples o Eucalyp us and Pine pulp suspension
lows a e epo ed he e, demons a ing he in luence o ib e concen a ion and ype, as well as low
eloci y, on he dis ibu ion o he ib es in he c oss-sec ional a ea o a ho izon al pipe.
The de eloped Elec ical Impedance Tomog aphy sys em was used o cha ac e ise he beha iou
o pine and eucalyp us ib e suspension low h ough a pipe a di e en ib e concen a ions—1%, 1.5%
and 2.35% (w/w)—and a a ying low eloci ies (be ween 0.5 and 3 m/s) based on he c oss-sec ion
images ob ained o o ally de eloped lows. Especially o he pine ib e suspensions, i was clea ly
possible o iden i y a plug o ma ion, which was p obably due o he geome y o he pine ib es, which
we e longe han he eucalyp us ones. Wi h espec o hei usage o he es ima ion o suspension
eloci y, eucalyp us ib e suspensions we e used in h ee di e en concen a ions—1%, 2% and 3%
(w/w)—and a eloci ies o be ween 0.25 and 1 m/s. The eal suspension eloci y measu ed by a
lowme e was qui e easonably es ima ed by he c oss co ela ion o EIT images ob ained in wo
consecu i e elec ode ings sepa a ed om one ano he by 300 mm.
The de eloped EIT sys em allows 2D image econs uc ion a es o up o 4000 o 1000 ames
pe second, depending on he numbe o elec odes used in he acquisi ion (16 o 32). Wi h espec o
he esolu ion o he ob ained c oss-sec ion images, i is much highe han p e iously epo ed in he
li e a u e (which is e iden by compa ing Figu e 3[53] and Figu e 8).
Thus a , he au ho s ha e used he de eloped equipmen no only in he moni o ing o ib e
suspension low and he es ima ion o ib e low eloci y, bu also wi h g ea success in o he di e se
wo-phase sys ems (solid–liquid [
72
], liquid–liquid [
111
], jus o men ion wo). In he nex s ep, he
objec i e will be o upg ade he o e all sys em (signal acquisi ion and image econs uc ion algo i hms)
o isualise h ee-phase sys ems.
Conside ing he esul s p esen ed he e, he use ulness o his echnique and equipmen is
demons a ed and clea ly jus i ies i s use a an indus ial scale, despi e i s lowe spa ial esolu ion
when compa ed wi h o he omog aphic p ocedu es.
Au ho Con ibu ions:
All au ho s con ibu ed equally o his wo k. All au ho s ha e ead and ag eed o he
published e sion o he manusc ip .
Funding:
This esea ch was unded by COST Ac ion FP1005, Fib e suspension low modelling—A key o
inno a ion and compe i i eness in he pulp & pape indus y, whose suppo is g a e ully acknowledged. The
au ho s hank he inancial suppo o Fundaç
ã
o pa a a Ci
ê
ncia e Tecnologia h ough he p ojec , con ac
Appl. Sci. 2020,10, 2355 24 o 30
PTDC/EQU-EQU/112388/2009 (COMPETE-01-0124-FEDER-015247), and con ac s Pes -C/EME/UI0285/2013 and
Pes -C/EQB/UI0102/2013, bo h inanced by FCT/MCTES (PIDDAC) and co- inanced by he Eu opean Regional
De elopmen Fund h ough he p og am COMPETE.
Con lic s o In e es : The au ho s decla e no con lic o in e es .
Re e ences
1.
S eine , G.; Podd, F. A non-in asi e and non-in usi e ul asonic ansduce a ay o p ocess omog aphy.
In P oceedings o he XVIII IMEKO WORLD Cong ess on Me olo y and Sus ainable De elopmen , Rio de
Janei o, B azil, 17–20 Sep embe 2006; pp. 17–22.
2.
Ijaz, U.Z.; Kim, J.-H.; Khambampa i, A.K.; Kim, M.-C.; Kim, S.; Kim, K.-Y. Concen a ion dis ibu ion
es ima ion o luid h ough elec ical impedance omog aphy based on in e ac ing mul iple model scheme.
Flow Meas. Ins um. 2007,18, 47–56. [C ossRe ]
3.
Mu phy, S.C.; Yo k, T.A. Elec ical impedance omog aphy wi h non-s a iona y elec odes. Meas. Sci. Technol.
2006,17, 3042–3052. [C ossRe ]
4.
Du s , F.; Melling, A.; Whi elaw, J.H. P inciples and P ac ice o Lase -Dopple Anemome y; Academic P ess:
New Yo k, NY, USA, 1981.
5.
Chaouki, J.; La achi, F.; Duduko ic, M.P. Non-in asi e omog aphic and elocime ic moni o ing o
mul iphase lows. Ind. Eng. Chem. Res. 1997,36, 4476–4503. [C ossRe ]
6.
Takeda, Y. Veloci y P o ile Measu emen by Ul asonic Dopple Shi Me hod. In P oceedings o he Na ional
Hea T ans e Con e ence o he Ame ican Socie y o Mechanical Enginee s, Philadelphia, PA, USA, 6–9
Augus 1989; pp. 155–160.
7.
Haa is o, S.; Ca dona, M.; Salmela, J.; Powell, R.; McCa hy, M.; Ka aja, M.; Koponen, A. Expe imen al
in es iga ion o he low p ope ies and heology o complex luids in pipe lows by hyb id mul iscale
elocime y. Exp. Fluids 2017,58, 158. [C ossRe ]
8.
So okin, M.; S okina, N.; Ee ola, T.; Lensu, L.; Ka unen, K.; Kal iainen, H. Image-based cha ac e isa ion o
he pulp lows. Pa e n Recogni . Image Anal. 2016,26, 630–637. [C ossRe ]
9.
Heindel, T. A e iew o X- ay low isualisa ion wi h applica ions o mul iphase lows. J. Fluids Eng.
2011
,
133. [C ossRe ]
10. Schliepe , G. P inciples o Gamma ay densi ome y. Me . Powde Rep. 2000,55, 20–23. [C ossRe ]
11.
Yenjaichon, W.; G ace, J.R.; Lim, C.J.; Benning on, C.P.J. Cha ac e isa ion o gas mixing in wa e and
pulp-suspension low based on elec ical esis ance omog aphy. Chem. Eng. J.
2013
,214, 285–297. [C ossRe ]
12.
Nakagawa, M.; Al obelli, S.A.; Cap ihan, A.; Fukushima, E.; Jeong, E.K. Non-in asi e measu emen s o
g anula lows by magne ic esonance imaging. Exp. Fluids 1993,16, 54–60. [C ossRe ]
13.
Sede man, A.J.; Gladden, L.F.; Man le, M.D. Applica ion o magne ic esonance imaging echniques o
pa icula e sys ems. Ad . Powde Technol. 2007,18, 23–38. [C ossRe ]
14.
Mülle , C.R.; Holland, D.J.; Sede man, A.J.; Man le, M.D.; Gladden, L.F.; Da idson, J.F. Magne ic esonance
imaging o luidized beds. Powde Technol. 2008,183, 53–62. [C ossRe ]
15.
Li, T.Q.; Ödbe g, L. Flow p ope ies o cellulose ibe suspensions loccula ed by ca ionic polyac ylamide.
Colloids Su . A Physiochem. Eng. Asp. 1996,115, 127–135. [C ossRe ]
16.
Li, T.Q.; Ödbe g, L. S udies o Floccula ion in Cellulose Fib e Suspensions by NMR Imaging. J. Pulp Pap. Sci.
1997,23, 401–405.
17.
Li, T.Q.; Powell, R.L.; Ödbe g, L.; McCa hy, M.J.; McCa hy, K.L. Veloci y measu emen s o ibe suspensions
by he nuclea magne ic esonance imaging me hod. Tappi J. 1994,77, 145–149.
18.
Li, T.Q.; Seymou , J.D.; Powell, R.L.; McCa hy, M.J.; McCa hy, K.L.; Ödbe g, L. Visualiza ion o low pa e ns
o cellulose ibe suspensions by NMR imaging. Aiche J. 1994,40, 1408–1411. [C ossRe ]
19.
Li, T.Q.; Weldon, M.; Ödbe g, L.; McCa hy, M.J.; Powell, R.L. Pipe low beha iou o ha dwood pulp
suspensions s udied by NMRI. J. Pulp Pap. Sci. 1995,21, 408–414.
20.
A ola, D.F.; Li, T.Q.; Ödbe g, L.; Powell, R.L.; McCa hy, M.J. Flow o pulp suspensions h ough a sudden
expansion. Aiche J. 1998,44, 2597–2606. [C ossRe ]
21.
Seymou , J.D.; Mane al, J.E.; McCa hy, K.L.; McCa hy, M.J.; Powell, R.L. NMR eloci y phase encoded
measu emen s o ib ous suspensions. Phys. Fluids A Fluid Dyn. 1993,5, 3010–3012. [C ossRe ]
Appl. Sci. 2020,10, 2355 25 o 30
22.
Sla ica, R. Lase Dopple Anemome y and i s Applica ion in Wind Tunnel Tes s. Sci. Tech. Re .
2007
,LVII,
64–76.
23.
Ek, R.; Mölle , K.; No man, B. Measu emen o eloci y and concen a ion a ia ions in dilu e ibe /ai
suspensions using a lase Dopple anemome e . Tappi J. 1978,61, 49–52.
24.
Ke ekes, R.J.; Ga ne , R.G. Measu emen o Tu bulence in Pulp Suspensions by Lase Anemome y. T ans.
Tech. Sec . 1982,CPPA 8, TR53.
25.
S een, M. The Applica ion o Re ac i e Index Ma ching o Two-phase Flow Measu emen s in Tu bulen
Fib e Suspensions by Lase Dopple Anemome y. No d. Pulp Pap. Res. J. 1989,4, 236–243. [C ossRe ]
26.
S een, M. On Tu bulence S uc u e in Ve ical Pipe Flow o Fibe Suspensions. No d. Pulp Pap. Res. J.
1989
,4,
244–252. [C ossRe ]
27.
Ande sson, S.; Rasmuson, A. Flow Measu emen s on a Tu bulen Fib e Suspension by Lase Dopple
Anemome y. Aiche J. 2000,46, 1106–1119. [C ossRe ]
28.
Pe e sson, J.; Rasmuson, A. LDA Measu emen s on a Tu bulen Gas/Liquid/Fib e Suspension. Can. J. Chem.
Eng. 2004,82, 265–274. [C ossRe ]
29.
Pe e sson, J.; Wiks öm, T.; Rasmuson, A. Nea Wall S udies o Pulp Suspension Flow Using LDA. Can. J.
Chem. Eng. 2006,84, 422–430. [C ossRe ]
30.
Takeda, Y. Veloci y P o ile Measu emen by Ul asonic Dopple Me hod. Exp. Fluid Sci.
1995
,10, 444–453.
[C ossRe ]
31.
Wiklund, J.; Pe e sson, J.; S ading, M.; Rasmuson, A. A Compa a i e S udy o UVP and LDA Techniques o
Pulp Suspensions in Pipe Flow. Aiche J. 2006,52, 484–495. [C ossRe ]
32.
Ozaki, Y.; Kawaguchi, T.; Takeda, Y.; Hishida, K.; Maeda, M. High ime esolu ion ul asonic eloci y p o ile .
Exp. Fluid Sci. 2002,26, 253–258. [C ossRe ]
33.
Hi simäki, O. De e mina ion o Radial Veloci y P o ile and Flow Dis u bance o Pulp Suspension by Ul asonic
Echo Co ela ion. Pap. Ja Puu 1978,60, 95–97.
34.
Ka ema, H.; Ka aja, M.; Kellomäki, M.; Salmela, J.; Selenius, P. T ansien Fluidisa ion o Fib e Suspension in
S aigh Channel Flow. In P oceedings o he TAPPI In e na ion Pape Physics Con e ence, San Diego, CA,
USA, 26–30 Sep embe 1999; pp. 369–379.
35.
Ka ema, H.; Salmela, J.; Tukiainen, M.; Lepomäki, H. P edic ion o Pape Fo ma ion by Fluidisa ion and
Re loccula ion Expe imen s. In P oceedings o he 12 h Fundamen al Resesea ch Symposium, Ox o d, UK,
11 Oc obe 2001; pp. 559–589.
36.
Wiklund, J.; Johansson, M.; Shaik, J.; Fische , P.; Windhab, E.; S ading, M.; He mansson, A.-M. In-Line
Ul asound based Rheome y o indus ial and model suspensions lowing h ough pipes. In P oceedings
o he Thi d In e na ional Symposium on Ul asonic Dopple Me hods o Fluid Mechanics and Fluid
Enginee ing, Lausanne, Swi ze land, 9–11 Sep embe 2002; pp. 69–76.
37.
Claesson, J.; Rasmuson, A.; Wiklund, J.; Wiks öm, T. Measu emen and analysis o low o concen a ed
ibe suspensions h ough sudden expansion using UVP. Aiche J. 2013,59, 1012–1021. [C ossRe ]
38.
Ko z
é
, R.; Wiklund, J.; Haldenwang, R. Op imisa ion o Pulsed Ul asonic Velocime y sys em and ansduce
echnology o indus ial applica ions. Ul asonics 2013,53, 459–469. [C ossRe ] [PubMed]
39.
Xu, H.; Aidun, C. Cha ac e is ics o ibe suspension low in a ec angula channel. In . J. Mul iph. Flow
2005
,
31, 318–336. [C ossRe ]
40.
D esxle , W.; Fujimo o, J. Op ical Cohe ence Tomog aphy: Technology and Applica ions; Sp inge : Be lin, Ge many, 2008.
41.
Ka aja, M.; Haa is o, S.; Slamela, J.; Le ho, R.; Koponen, A. Cha ac e isa ion o mic o- ib ila ed cellulose
ibe suspension low using mul i scale eloci y p o ile measu emen s. No d. Pulp Pap. J.
2017
,31, 473–482.
[C ossRe ]
42.
Fujimo o, J.; Pis is, C.; Boppa , S.; B ezinski, M. Op ical cohe ence omog aphy: An eme ging echonology
o biomedical imaging and op ical biopsy. Neoplasia 2000,2, 9–25. [C ossRe ]
43.
Gao, Q.; Wang, H.; Shen, G. Re iew on de elopmen o olume ic pa icle image elocime y. Chin. Sci.
Bull. 2013,58, 4541–4556. [C ossRe ]
44.
Fan, L.; Xu, N. S udy on lamina low con aining ib e pa icles in a s i ed ank using DPI. Powde Technol.
2017,320, 420–427. [C ossRe ]
45.
Fock, H.; Claesson, J.; Rasmusson, A.; Wiks öm, T. Nea wall e ec in he plug low o pulp suspensions.
Can. J. Chem. Eng. 2011,89, 1207–1216. [C ossRe ]