Transporting Highly Concentrated Slurries With Centrifugal Pumps

L I C E N T I AT E T H E S I S
Thord Wennberg Transporting Highly Concentrated Slurries with Centrifugal Pumps The Thickened Minerals Tailings Example
Department of Chemical Engineering and Geosciences
Division of Mineral Processing
ISSN: 1402-1757 ISBN 978-91-7439-169-5

Transporting Highly Concentrated Slurries
Luleå University of Technology 2010 with Centrifugal Pumps
The Thickened Minerals Tailings Example
Thord Wennberg
Transporting Highly Concentrated Slurries
with Centrifugal Pumps
The Thickened Minerals Tailings Example
Thord Wennberg
Luleå University of Technology

Department of Chemical Engineering and Geosciences
Mineral Processing
Printed by Universitetstryckeriet, Luleå 2010
ISSN: 1402-1757
ISBN 978-91-7439-169-5
Luleå 2010
www.ltu.se
Preface
7KLVWKHVLVLVSUHVHQWHGDVDSDUWLDOIXOILOPHQWRIWKHUHTXLUHPHQWVIRUWKHGHJUHHRI/LFHQWLDWH
DW/XOHn8QLYHUVLW\RI7HFKQRORJ\,WFRQFHUQVWKHK\GUDXOLFWUDQVSRUWRIKLJKO\FRQFHQWUDWHG
VOXUULHVZLWKFHQWULIXJDOSXPSVLQFRQQHFWLRQZLWKWKLFNHQHGPLQHUDOVWDLOLQJVV\VWHPV
,ZDQWWRH[SUHVVP\WKDQNVWRP\H[DPLQHU'RFHQW%HUWLO3nOVVRQDWWKH'LYLVLRQRI0LQHUDO
3URFHVVLQJDQGVXSHUYLVRU3URIHVVRU(PHULWXV$QGHUV6HOOJUHQIRUSURYLGLQJFRQVWUXFWLYHDQG
YDOXDEOHJXLGDQFH7KHLUFRQWULEXWLRQDQGSDWLHQFHZLWKPHWKURXJKRXWWKHVHVWXGLHVLVJUHDWO\
DSSUHFLDWHG

, ZRXOG DOVR OLNH WR H[SUHVV P\ WKDQNV WR P\ FROOHDJXHV DW /XRVVDYDDUD.LLUXQDYDDUD $%
/.$%&RPSDQ\IRUVXSSRUWHVSHFLDOO\Dr.cVD6XQGTYLVWIRUKHUJXLGDQFHLQWKHLQLWLDO
SKDVH

0\DSSUHFLDWLRQDOVRJRHVWR-DQ$QGHUVVRQDQG2YH(ULNVVRQ0HWVR0LQHUDOV6ZHGHQ$%
IRUWHFKQLFDOVXSSRUWDQGLQIRUPDWLRQ,ZRXOGDOVROLNHWRH[SUHVVP\WKDQNVWR1HZ%ROLGHQ
DQG WKH SHUVRQQHO DW WKH SDVWH SODQW DW WKH *DUSHQEHUJ PLQHV IRU RSHUDWLQJ GDWD IURP D
WKLFNHQHUXQGHUIORZSXPSLQVWDOODWLRQ

)LQDOO\,WKDQNIXOO\DFNQRZOHGJH/.$%IRUJLYLQJPHWKHRSSRUWXQLW\WRSHUIRUPWKLVZRUN

7KRUG:HQQEHUJ

0DOPEHUJHW1RYHPEHU

ABSTRACT

1HZ WHFKQRORJLHV DUH FRQWLQXRXVO\ FRQVLGHUHG IRU WKH KDQGOLQJ DQG GHSRVLWLRQ RI WDLOLQJV
:LWKWKLFNHQHGWDLOLQJVRQO\DIUDFWLRQRIZDWHUXVHGFRQYHQWLRQDOO\UHDFKHVWKHGLVSRVDODUHD
WKXVOLPLWLQJVL]HVRIZDWHUKROGLQJGDPIDFLOLWLHVDQGDUHDVWREHUHKDELOLWDWHG+HDWPD\DOVR
EH UHFRYHUHG ZLWK GLUHFW ZDWHU FLUFXODWLRQ IURP WKH WKLFNHQLQJ 7HFKQLFDOHFRQRPLFDO
IHDVLELOLW\ FRQVLGHUDWLRQV DUH KHUH FRXSOHG WR WKH 6YDSSDYDDUD LURQ RUH RSHUDWLRQV E\ WKH
/XRVVDYDDUD.LLUXQDYDDUD$%/.$%&RPSDQ\LQQRUWKHUQ6ZHGHQ/RFDOFRQGLWLRQVKHUH
IDYRXUSODFHPHQWZLWKDQHOHYDWHGWKLFNHQHUORFDWLRQFORVHWRWKHGLVSRVDODUHD

7KHWKLFNHQHGWDLOLQJVVWXGLHGKHUHKDYHW\SLFDOO\DYHUDJHSDUWLFOHVL]HVRIWRPZLWK
PD[LPXPVL]HVRIXSWRDERXWP7KHFRUUHVSRQGLQJVOXUULHVFRYHUDQLQWHUPHGLDWHDUHD
EHWZHHQKRPRJHQHRXVO\DQGKHWHURJHQHRXVO\IORZLQJPL[WXUHV9DULRXVGHVLJQIHDWXUHVDUH
RIWHQ UHIHUHHG WR LQ UKHRORJLFDO WHUPV LH KRPRJHQHRXV IORZ +HUH WKHVH WHUPV DUH
FRQVLGHUHGZLWKVRPHFDXWLRQGXHWRWKHQRWDEOHDPRXQWRISDUWLFOHVODUJHUWKDQDERXWP

7KHREMHFWLYHLVWRIRUPDEDFNJURXQGIRUIHDVLELOLW\FRQVLGHUDWLRQVEDVHGRQODERUDWRU\DQG
SLORWVFDOHH[SHULPHQWVLQSLSHOLQHVDQGIOXPHVWRJHWKHUZLWKYLVFRPHWULFPHDVXUHPHQWVDQG
GHSRVLWLRQDOVORSHREVHUYDWLRQV3LSHOLQHGLDPHWHUVFRYHUHGWKHIXOOVFDOHUDQJHXSWRP

,WZDVIRXQGWKDWDVROLGVFRQFHQWUDWLRQE\YROXPHRIDERXWFKDUDFWHUL]HVWKHSLSHOLQH
SXPSLQJ KHUH HVWLPDWHG WR EH RQ WKH FRQVHUYDWLYH VLGH ZLWK UHVSHFW WR WKLFNHQHG SURGXFW
TXDOLW\ UHTXLUHPHQW 7KH RYHUDOO SLSHOLQH ORRS UHVXOWV ZHUH HYDOXDWHG LQ WHUPV RI D
UHSUHVHQWDWLYH SLSH ZDOO VWUHVV RI DERXW 3D $SSUR[LPDWH YDOXHV RI WKH %LQJKDP \LHOG
VWUHVVDQGSODVWLFYLVFRVLW\ZHUHHVWLPDWHGWR3DDQG3DVUHVSHFWLYHO\'HSRVLWLRQDO
VORSHVKDYHEHHQHVWLPDWHGWRDERXWIRUVOXUU\SURSHUWLHVWKDWJLYHVDFRQFHSWXDOO\HYHQ
LQFOLQDWLRQZLWKQRVHJUHJDWLRQRISDUWLFOHVDQGYLUWXDOO\QRGUDLQDJHRIZDWHU

:LWKDQHOHYDWHGWKLFNHQHUORFDWLRQFORVHWRWKHGLVSRVDODUHDDQHZKLJKSUHVVXUHSXPSLQJ
V\VWHPIRUDERXW03DLVDYRLGHGZKLOHWKHWKLFNHQHUXQGHUIORZVOXUU\LVGHOLYHUHGZLWKD
FHQWULIXJDOSXPSDVKRUWGLVWDQFHGXULQJWKHILUVW\HDUVRIRSHUDWLRQ,QRUGHUWRPHHWWKHPRVW
HIIHFWLYH ORQJWHUP VROXWLRQ IRU SLSHOLQH GLVWULEXWLRQ DQG SODFHPHQW FRQWLQXHG XVH RI
FHQWULIXJDOSXPSVLQVHULHVLQVWDOODWLRQVLVFRQVLGHUHGDVWKHGLVFKDUJHSRLQWDGYDQFHVGXULQJ
WKH\HDUOLIH

7KHSHUIRUPDQFHRIDFHQWULIXJDOSXPSLVDIIHFWHGZKHQKDQGOLQJKLJKO\FRQFHQWUDWHGVOXUULHV
ZKLFKVKRZVXSDVUHGXFWLRQVRIWKHFOHDUZDWHUKHDGDQGHIILFLHQF\FXUYHV,WZDVIRXQGWKDW
PD[LPXP UHGXFWLRQV LQ KHDG DQG HIILFLHQF\ ZHUH DERXW DQG UHVSHFWLYHO\ ZKHQ
RSHUDWLQJ DW WKH GHVLJQ IORZ UDWH 4%(3 LQ WKH EHVW HIILFLHQF\ UHJLRQ $V RSSRVHG WR
1HZWRQLDQIOXLGVDGLYHUJLQJWHQGHQF\IURPWKHZDWHUKHDGFXUYHIRUGHFUHDVLQJIORZUDWHV
KDV VRPHWLPHV EHHQ UHSRUWHG EHORZ DERXW 4%(3 IRU VOXUULHV WKDW VKRZ D KLJKO\ QRQ
1HZWRQLDQ EHKDYLRXU 9DULRXV PHFKDQLVPV UHODWHG WR WKH GLYHUJLQJ WHQGHQF\ DUH GLVFXVVHG
7KHVXFWLRQVLGHSUHVVXUHUHTXLUHPHQWLVQRUPDOO\IXOILOOHGZLWKDODUJHPDUJLQIRUDWKLFNHQHU
XQGHUIORZSXPS,WZDVREVHUYHGLQDQLQVWDOODWLRQKRZDVWDQGDUGSXPSRSHUDWLQJDW4%(3
XQGHUDWKLFNHQHUVKRZHGQRXQVWDEOHKHDGWHQGHQFLHVIRUDWDLOLQJVVOXUU\ZLWKDQ\LHOGVWUHVV
RI QHDUO\ 3D +RZHYHU SXPSV LQ VHULHV LQVWDOODWLRQ VKRXOG EH DUUDQJHG ZLWK VXIILFLHQW
PDUJLQVIRUDQLQFUHDVHGVXFWLRQVLGHSUHVVXUHUHTXLUHPHQW

.H\ZRUGV7DLOLQJVVOXUU\KLJKO\FRQFHQWUDWHGQRQ1HZWRQLDQEHKDYLRXUFHQWULIXJDOSXPS
SHUIRUPDQFHWDLOLQJVGHSRVLWLRQDOFRQGLWLRQV

Published papers
A
:HQQEHUJ76HOOJUHQ$:KLWORFN/3UHGLFWLQJWKHSHUIRUPDQFHRIFHQWULIXJDO
SXPSVZKHQKDQGOLQJFRPSOH[VOXUULHV3URFHHGLQJVWK,QWHUQDWLRQDO&RQIHUHQFHRQ
7UDQVSRUW 6HGLPHQWDWLRQRI6ROLGVSDUWLFOHV-XQH6DLQW3HWHUVEXUJ5XVVLDSS

B
:HQQEHUJ76HOOJUHQ$3XPSLQJHYDOXDWLRQVZLWKSDVWHWDLOLQJVWKLFNHQHGFORVHWR
WKHVXUIDFHGLVSRVDODUHD3URFHHGLQJV7KHWKLQWHUQDWLRQDOFRQIHUHQFHRQWKHK\GUDXOLF
WUDQVSRUWDWLRQRIVROLGVLQSLSHV+\GURWUDQVSRUW&DSH7RZQ6RXWK$IULFD0D\SS

C
:HQQEHUJ76HOOJUHQ$*ROGNXKO,5KHRORJLFDODQG'HSRVLWLRQDOFKDUDFWHULVDWLRQ
RISDVWHOLNHWDLOLQJVVOXUULHV3URFHHGLQJV&RQIHUHQFHLQ0LQHUDOV(QJLQHHULQJ)HEUXDU\
/XOHn6ZHGHQ

D
:HQQEHUJ76HOOJUHQ$6XQGTYLVWc)HDVLELOLW\FRQVLGHUDWLRQVRIKLJKGHQVLW\
GLVSRVDODWWKH/.$%6YDSSDYDDUDFRQFHQWUDWRU3URFHHGLQJV7KH,QW&RQIHUHQFH6HFXULQJ
WKH)XWXUH-XQH6NHOOHIWHn6ZHGHQ

E
(QJPDQ0$6HOOJUHQc6XQGTYLVW,*ROGNXKODQG7:HQQEHUJ8VHUV
SHUVSHFWLYHRQWKHGHVLJQRIKLJKGHQVLW\EDVHPHWDOWDLOLQJVV\VWHPV3URFHHGLQJVWK,QW
&RQIHUHQFHRQ7DLOLQJVDQG0LQH:DVWH¶9DLO&286$2FWREHU

Contents

ABSTRACT .............................................................................................................................. 5

PUBLISHED PAPERS ............................................................................................................ 7

CONTENTS .............................................................................................................................. 9

INTRODUCTION .................................................................................................................. 11
&RQYHQWLRQDODQGWKLFNHQHGWDLOLQJVPDQDJHPHQW

2. OBJECTIVES AND SCOPE............................................................................................. 15

3. FLOW OF THICKENED TAILINGS SLURRIES ........................................................ 17
&KDUDFWHUL]DWLRQ
&RQGXLWIORZUHVLVWDQFH
)UHHVXUIDFHIORZ

4. PERFORMANCE OF CENTRIFUGAL PUMPS HANDLING COMPLEX
SLURRIES .............................................................................................................................. 21

5. EXPERIMENTAL RESULTS .......................................................................................... 23
)ORZUHVLVWDQFHLQSLSHOLQHVDQGIOXPHV
'HSRVLWLRQVORSHHYDOXDWLRQV
3XPSSHUIRUPDQFH
2YHUDOOUHVXOWVXPPDU\

6. DISCUSSION ..................................................................................................................... 27
&HQWULIXJDOSXPSSHUIRUPDQFH
'HSRVLWLRQDOFRQGLWLRQV

7. CONCLUSIONS................................................................................................................. 35

8. FUTURE WORK ............................................................................................................... 37

9. REFERENCES ................................................................................................................... 39

10. NOTATIONS.................................................................................................................... 43

1. Introduction

7KHUH LV DQ LQFUHDVHG LQWHUHVW WKURXJKRXW WKH ZRUOG LQ WKH WUHDWPHQW DQG SURFHVVLQJ RI ILQH
JUDLQHGUHViGuDO SURGXFWV³WDLOLQJV´IURPPLQHUDO EHQHILFLDWLRQ SODQWVOLPLWWKHYROXPHVWR
EHKDQGOHGDQGUHGXFHVL]HVRIWDLOLQJGDPV 7KHHFRQRPLFDO IHDVLELOLW\ RIDKLJKGHJUHHRI
WDLOLQJV WKLFNHQLQJ LV GHWHUPLQHG E\ WKH ORQJWHUP LPSDFW RQ WKH GLVSRVDO DUHD :DWHU
DYDLODELOLW\ PD\ EH D FULWLFDO LVVXH LQ DULG RU VHPLDULG DUHDV ZKHUH ZDWHU H[SRVHG LQ WKH
GLVSRVDODUHDLVODUJHO\ORVWWKURXJKHYDSRUDWLRQ,QVXEDUFWLFDQGDUFWLFUHJLRQVKHDWUHFRYHU\
ZLWK UDSLG ZDWHUFLUFXODWLRQ WKURXJKWKLFNHQLQJFDQLPSURYHWKH HQHUJ\ HIIHFWLYHQHVV LQ WKH
PLQHUDOSURFHVVLQJSODQW

7KHUHGXFHGFRVWIRUGDPVFDSSLQJDQGUHWXUQZDWHUSXPSLQJFDQEHFRQVLGHUDEOHRYHUWKH
OLIHRIWKHPLQH$QRWKHUFRQFHUQLVGDPVDIHW\VRPHUHSRUWHGWDLOLQJVGDPIDLOXUHVDQG
LQFLGHQWVPD\EHUHODWHGWRODUJHDPRXQWVRIZDWHULQWKHGLVSRVDODUHD
1.1 Conventional and thickened tailings management
7KHWDLOLQJVVOXUU\LVFRQYHQWLRQDOO\GLVFKDUJHGZLWKDFRQVLGHUDEOHTXDQWLW\RISURFHVVZDWHU
LQWRDVHWWOLQJSRQG)LJ

Pipeline Dam
Tailings Pond

Figure 1. Conventional tailings disposal system.

,Q JHQHUDO VORSHV DUH EHORZ LQ WKH GLVSRVDO DUHD 0RVW RI WKH SURFHVV ZDWHU LV WKHQ
UHFRYHUHG IURP D VHFRQG SRQG FODULILFDWLRQ SRQG DQG UHWXUQHG WR WKH SURFHVV 7KH VROLGV
FRQFHQWUDWLRQ LV RIWHQ ZHOO EHORZ E\ YROXPH E\ PDVV ZLWK D SRWHQWLDO RI
VHJUHJDWLRQDIWHUGHSRVLWLRQLHK\GUDXOLFVRUWLQJRISDUWLFOHVLQWKHVHWWOLQJSRQG
:LWKDWDLOLQJVV\VWHPWKDWLQFOXGHVDWKLFNHQHUWKDWSURGXFHVDKLJKO\FRQFHQWUDWHGXQGHUIORZ
WKH WKLFNHQHU FDQ EH SODFHG QHDU WKH FRQFHQWUDWRU RU LQ WKH GLVSRVDO DUHD ,Q IODW WHUUDLQ WKH
WKLFNHQHGWDLOLQJVLVGLVFKDUJHGIURPDQDUWLILFLDOUDPSRUDWRZHUIRUPLQJDULGJHRUFRQH,W
PD\ EH IDYRXUDEOH WR ORFDWH WKH WKLFNHQHU FORVH WR WKH GLVSRVDO DUHD ZKHQ GHSRVLWLRQ WDNHV
SODFHDORQJDKLOOVLGHRUGRZQDYDOOH\)LJXUH
7KLFNHQHU
&RQFHQWUDWRU
3XPS
:DWHU 'DP
3XPS
Figure 2. Schematic descriptions of a thickened tailings system with the thickener located
on a hill close to the disposal area.

:LWK DQ HOHYDWHG ORFDWLRQ RI WKH WKLFNHQHU FORVH WR WKH GLVSRVDO DUHD RQO\ VKRUW GLVWDQFH
SXPSLQJRIWKLFNHQHGVOXUU\LVLQLWLDOO\UHTXLUHG:LWKDWKLFNHQHGWDLOLQJVV\VWHPXSWR
RI WKH ZDWHU FDQ EH UHF\FOHG GLUHFWO\ IURP WKH RYHUIORZ RI WKH WKLFNHQHU 7KH VROLGV
FRQFHQWUDWLRQVE\YROXPH&PXVWJHQHUDOO\EHZHOODERYHWRDFKLHYHVOXUU\SURSHUWLHV
WKDW JLYHV D FRQFHSWXDOO\ HYHQ VORSH RI GHSRVLWHG WDLOLQJV RI ZLWK QR VHJUHJDWLRQ RI
SDUWLFOHVDQGYLUWXDOO\QRGUDLQDJHRIZDWHU

7KHVOXUU\GHQVLW\UDWLRLVUHODWHGWR&LQWKHIROORZLQJZD\
1 C S S 1
O

ZKHUH oLVWKHGHQVLW\RIZDWHUDQG6VWKHVROLGVGHQVLW\UDWLR
SS S

O

ZKHUH SLVWKHGHQVLW\RIVROLGV)RUH[DPSOHZLWK& DQGSs WKHVOXUU\GHQVLW\
UDWLRLQ(TLVDERXW

7KH UHODWLRQVKLSV EHWZHHQ WKH VROLGV FRQFHQWUDWLRQ E\ YROXPH DQG PDVV IORZ UDWH RI VROLGV
DQG ZDWHU DUH SUHVHQWHG LQ )LJXUH WRJHWKHU ZLWK D JUDSKLFDO UHSUHVHQWDWLRQ RI WKH ZDWHU
UHTXLUHPHQWSHUWRQQHRIGU\VROLGV

Cw
C
Ss Cw (Ss 1)

Ms
Cw
Tonne of solids
Ms Mo
Tonne water

Mo 1 Cw
Ms Cw

Solids concentration by mass

Solids concentration by volume
Figure 3. The relationship between water content per tonne solids and solids concentration
or the volume metric concentration, Ss = 2.9. Ms= mass flow rate of dry solids,
MO= mass flow rate of water. Cw is concentration by mass.

,W IROORZV IURP )LJXUH KRZ WKH ZDWHU UHTXLUHPHQW LV VWURQJO\ QRQOLQHDUO\ UHODWHG WR WKH
VROLGV FRQFHQWUDWLRQV ,Q WKH FDVH RI WKLFNHQHG WDLOLQJ WKH UHFLUFXODWLRQ RI SURFHVV ZDWHU LV
UHFRYHUHG GLUHFWO\ IURP WKH WKLFNHQHU WR WKH FRQFHQWUDWRU LQVWHDG RI IORZLQJ WKURXJK WKH

GLVSRVDODUHD7KHUHFRYHU\RIKHDWLQDWKLFNHQHGWDLOLQJVV\VWHPPD\EHDYDOXDEOHUHVRXUFH
LQFROGDUHDVVHHIRUH[DPSOH+ROORZHWDO2[HQIRUGDQG/RUG

$FRPSDULVRQRIWKHIORZVRIZDWHUDQGVOXUU\LQDFRQYHQWLRQDODQGWKLFNHQHGWDLOLQJVV\VWHP
LVVKRZQLQ7DEOHWRJHWKHUZLWKDQLQGLFDWLRQRIWKHKHDWUHFRYHU\SRWHQWLDO

Table 1. Flows of slurry and water and heat recovery potential in an example with the
capacity 90 tonnes /h of dry tailings in a sub-arctic region.
&RQYHQWLRQDO 7KLFNHQHGWDLOLQJV
6ROLGVFRQFHQWUDWLRQE\PDVV
)ORZRIVOXUU\PK
)ORZRIZDWHUPK
5HFRYHUHGZDWHUWHPSHUDWXUHR&
$YHUDJHWHPSHUDWXUHRYHUWKH\HDU

7KHORFDWLRQRIWKHWKLFNHQHULVLQDGGLWLRQWRORFDOFRQGLWLRQVPDLQO\DEDODQFHRIWKHFRVWIRU
KLJKSUHVVXUHSXPSLQJDQGWKHGLVDGYDQWDJHRIKDYLQJWKHWKLFNHQLQJIDFLOLW\UHPRWHIURPWKH
UHVW RI WKH SURFHVVLQJ WDNLQJ SODFH LQ WKH FRQFHQWUDWRU 7HFKQLFDOHFRQRPLFDO IHDVLELOLW\
FRQVLGHUDWLRQVDUHFRXSOHGWRWKH6YDSSDYDDUDLURQRUHPLQHRSHUDWLRQVE\WKH/XRVVDYDDUD
.LLUXQDYDDUD$%/.$%&RPSDQ\LQQRUWKHUQ6ZHGHQ:HQQEHUJHWDO3UHOLPLQDU\
V\VWHP HYDOXDWLRQV KDG LQGLFDWHG WKDW WKH ORFDO FRQGLWLRQV IDYRXUHG SODFHPHQW ZLWK DQ
HOHYDWHGWKLFNHQHUORFDWLRQFORVHWRWKHGLVSRVDODUHD

:LWKWKHWKLFNHQHUORFDWHGDWWKHFRQFHQWUDWRULQYHVWPHQWVLQDKLJKSUHVVXUHV\VWHPZRXOGEH
UHTXLUHG ZLWK SRVLWLYH GLVSODFHPHQW W\SH RI SXPSV :LWK D ORFDWLRQ RQ WKH KLOO FHQWULIXJDO
SXPSV FDQ EH XVHG GLUHFWO\ GXULQJ WKH ILUVW \HDUV RI RSHUDWLRQ &HQWULIXJDO SXPSV DUH RIWHQ
XVHGWRSXPSKLJKGHQVLW\WKLFNHQHUXQGHUIORZVKRUWGLVWDQFHVRUWRFLUFXODWHWKLFNHQHGVOXUU\
LQWKHORZHUSRUWLRQRIWKHWKLFNHQHUWRSURPRWHPL[LQJDQGIORZDELOLW\

7R PHHW WKH PRVW HIIHFWLYH ORQJWHUP VROXWLRQ IRU SLSHOLQH GLVWULEXWLRQ DQG SODFHPHQW
FRQWLQXHGXVHRIFHQWULIXJDOSXPSVLQVHULHVLQVWDOODWLRQVLVKHUHFRQVLGHUHGDVWKHGLVFKDUJH
SRLQW DGYDQFHV GXULQJ WKH \HDU OLIH 7KH WKLFNHQHU RSHUDWLRQ SLSHOLQH GLVWULEXWLRQ DQG
SODFHPHQWRIKLJKO\WKLFNHQHGWDLOLQJVIRUPDV\VWHPZKLFKVKRXOGEHGHVLJQHGDQGPDQDJHG
IRUPD[LPXPORQJWHUPHIIHFWLYHQHVV

:LWK WKH KLJK UHTXLUHG YROXPHWULF VROLGV FRQFHQWUDWLRQV RI ZHOO DERYH WKH IORZ
UHVLVWDQFH LQ WKH WKLFNHQLQJ SXPSLQJ DQG GHSRVLWLRQ EH\RQG D FHUWDLQ FRQFHQWUDWLRQ ULVHV
GUDPDWLFDOO\IRULQFUHPHQWDOLQFUHDVHVLQWKHVROLGVFRQWHQW

2. Objectives and scope

7KH REMHFWLYH RI WKLV VWXG\ LV WR IRUP D EDFNJURXQG IRU IHDVLELOLW\ FRQVLGHUDWLRQV EDVHG RQ
ODERUDWRU\ DQG SLORWVFDOH H[SHULPHQWV LQ IOXPHV DQG SLSHOLQHV WRJHWKHU ZLWK YLVFRPHWULF
PHDVXUHPHQWV DQG VORSH REVHUYDWLRQV UHODWHG WR SLSHOLQH GLVWULEXWLRQ DQG SODFHPHQW
7KLFNHQLQJ ODERUDWRU\ DQG SLORWVFDOH WHVWV ZHUH FDUULHG RXW DQG UHSRUWHG VHSDUDWHO\ E\
HTXLSPHQWVXSSOLHUV

5HVXOWLQJ EDFNJURXQG IORZUHODWHG ODERUDWRU\ DQG SLORW VFDOH H[SHULPHQWV FDUULHG RXW LQ
FRQQHFWLRQ ZLWK WKLFNHQLQJ WHVWV GRQH E\ RWKHUV DUH SUHVHQWHG LQ 3DSHU ( WRJHWKHU ZLWK D
GLVFXVVLRQ RI VORSH REVHUYDWLRQV DQG SUHGLFWLRQV 7KLV GLVFXVVLRQ LV H[WHQGHG LQ 3DSHU &
ZKHUH VRPH UHSRUWHG LQVWDOODWLRQV DOVR DUH UHYLHZHG ,Q 3DSHU ' WKH WHFKQLFDOHFRQRPLFDO
IHDVLELOLW\LVEULHIO\SUHVHQWHGIRUWKHYDULRXVDUUDQJHPHQWVDQGWKLFNHQHUORFDWLRQVWKDWKDYH
EHHQFRQVLGHUHG([SHULPHQWDOSLSHOLQHUHVLVWDQFHUHVXOWVLQSLORWVFDOHDUHSUHVHQWHGLQ3DSHU
%WRJHWKHUZLWKDVXPPDUL]DWLRQDQGGLVFXVVLRQRIUHSRUWHGSXPSSHUIRUPDQFHGDWD3DSHU$
LVFRQFHQWUDWHGWRDGLVFXVVLRQRISXPSSHUIRUPDQFHUHVXOWVDQGPHFKDQLVPVZKHQKDQGOLQJ
FRPSOH[VOXUULHV

3. Flow of thickened taili ngs slurries
3.1 Characterization
7KHWDLOLQJVVWXGLHGKHUHKDYHW\SLFDOO\DYHUDJHSDUWLFOHVL]HVRIWRPZLWKPD[LPXP
VL]HV RI XS WR DERXW P 7KH FRUUHVSRQGLQJ VOXUULHV DUH WHUPHG FRPSOH[ EHFDXVH WKH\
FRYHU DQ LQWHUPHGLDWH DUHD EHWZHHQ KRPRJHQHRXVO\ DQG KHWHURJHQHRXVO\ IORZV RU ´QRQ
VHWWOLQJ´DQG³VHWWOLQJ´W\SHVRIPL[WXUHVUHVSHFWLYHO\,WLVNQRZQWKDWWKHVHVOXUULHVPD\QRW
EHKDYH LQ D FRPSOHWHO\ QRQVHWWOLQJ ZD\ IORZLQJ KRPRJHQHRXVO\ GXULQJ SLSHOLQH
WUDQVSRUWDWLRQEHFDXVHODUJHUSDUWLFOHVPD\VKRZVHWWOLQJWHQGHQFLHV3XOOXP

9DULRXV WKLFNHQHG WDLOLQJV GHVLJQ IHDWXUHV DUH RIWHQ UHIHUHHG WR LQ UKHRORJLFDO WHUPV LH
KRPRJHQHRXV IORZ 7KHVH WHUPV DUH XVHG KHUH ZLWK VRPH FDXWLRQ ZLWK UHVSHFW WR WKDW D
FRQVLGHUDEOHDPRXQWRISDUWLFOHVDUHODUJHUWKDQDERXWP

3.2 Conduit flow resistance
7KHSUHVVXUHJUDGLHQW p / x 3DPIRUWKHIORZLQDKRUL]RQWDOSLSHOLQHLV

p
g j
x

ZKHUHjLVWKHIULFWLRQORVVHVLQPRIVOXUU\SHUPRISLSHK\GUDXOLFJUDGLHQW LVWKHGHQVLW\
RIWKHVOXUU\gLVWKHDFFHOHUDWLRQGXHWRJUDYLW\PV7KHSLSHZDOOVKHDUVWUHVV wLV
UHODWHG WR WKH SUHVVXUH JUDGLHQW DQG WKH IULFWLRQ ORVV JUDGLHQW WKURXJK D IRUFH EDODQFH LQ D
SLSHOLQHVHFWLRQZLWKGLDPHWHUD

p x D gjD
gj
w
4 4

7KLV UHODWLRQ KROGV IRU ODPLQDU DQG WXUEXOHQW IORZV DV ZHOO DV IRU 1HZWRQLDQ RU QRQ
1HZWRQLDQIOXLGVDQGWKHSLSHZDOOVKHDUVWUHVVLVDNH\VFDOLQJSDUDPHWHU

7KH IORZ RI KLJKO\ FRQFHQWUDWHG WDLOLQJV VOXUULHV EHKDYHV VHHPLQJO\ OLNH KRPRJHQHRXVO\
IORZLQJQRQ1HZWRQLDQPHGLD5KHRORJLFDOGHILQLWLRQVDUHLQWURGXFHGEHORZIRUH[SUHVVLRQV
RIIORZUHVLVWDQFHDQGVFDOLQJ

7KHYLVFRVLW\ RID1HZWRQLDQIOXLGLVGHILQHGE\WKHUDWLRRIVKHDUVWUHVVĲDQGVKHDUUDWH
Ȗ VWDUWLQJ IURP ]HUR LH D OLQHDU UHODWLRQVKLS VHH )LJXUH $ %LQJKDP W\SH RI D QRQ
1HZWRQLDQIOXLGLVGHILQHGE\DSODVWLFYLVFRVLW\ SDQGD\LHOGVWUHVVĲ\WKDWPXVWEHH[FHHGHG
EHIRUHWKHIOXLGVWDUWVWRIORZ

y p

W
Pa
tic
plas
Apparent viscosity
Shear stress W
ham
Bing
PS
an
ni
P
to
W\
ew
N
Pf
Shear rate J J Shear rate J

Figure 4. Rheogram for a non-Newtonian Bingham plastic fluid and a Newtonian fluid
(left). The apparent viscosity a versus shear rate for a Bingham plastic fluid is
shown to the right together with a constant viscosity; μ∞, at high shear rates

,W LV VHHQ LQ )LJXUH KRZ WKH DSSDUHQW YLVFRVLW\ LV GHILQHG E\ WKH SLSH ZDOO VKHDU VWUHVV
GLYLGHG ZLWK WKH VKHDU UDWH DW DQ DUELWUDU\ SRLQW Ĳ Ȗ LH ȝD DW D SDUWLFXODU UDWH RI VKHDU
H[SUHVVHGLQWHUPVRID1HZWRQLDQIOXLG:LWKDQRQ1HZWRQLDQIOXLGZLWKD\LHOGVWUHVVWKH
DSSDUHQWYLVFRVLW\LQFUHDVHVGUDPDWLFDOO\IRUVPDOOVKHDUUDWHV$VWKHVKHDUUDWHLVLQFUHDVHG
WKH DSSDUHQW YLVFRVLW\ UHDFKHV DQ DOPRVW FRQVWDQW YDOXH WKDW FDQ EH FRPSDUHG ZLWK WKH
SODVWLFYLVFRVLW\LQ )LJXUH OHIW$FRQWLQXRXVO\ GHFUHDVLQJFXUYHEHWZHHQWKH1HZWRQLDQ
DQG%LQJKDPFXUYHVLQ)LJXUHVWDUWLQJIURP]HURGHILQHVDSVHXGRSODVWLFIOXLG,WVDSSDUHQW
YLVFRVLW\DOVRLQFUHDVHVZLWKGHFUHDVLQJVKHDUUDWH

7KH VKHDU VWUHVV DW WKH SLSH ZDOO Z DQG WKH SVHXGR VKHDU UDWH 8V/D UHSUHVHQW UKHRORJLFDO
VFDOLQJ SDUDPHWHUV ZKHUH 9 LV YHORFLW\ 5DELQRZLWFK DQG 0RRQH\ IRXQG D
WHFKQLTXHZKLFKUHODWHVWKHSVHXGRVKHDUUDWHWRWKHWUXHVKHDUUDWH

3n+1/4n8V/D

ZKHUHnUHSUHVHQWVWKHVORSHLQDORJORJSORWRIĲZYHUVXV9'

,Q WKLV ZD\ D UKHRJUDP FDQ EH FRQVWUXFWHG IURP SUHVVXUH GURS (TV DQG IORZ UDWH
PHDVXUHPHQWV RQ D SRLQWE\SRLQW EDVLV ZLWKRXW DQ\ UKHRORJLFDO PRGHO DSSURDFK :LWK WKH
%LQJKDPPRGHOLQ(TWKHIRUPDOUHODWLRQVKLSIRUWKHSVHXGRVKHDUUDWHFDQEHH[SUHVVHGDV
IROORZVGHWDLOVRPLWWHGKHUHVHHIRUH[DPSOH6KRRNDQG5RFR

4
8V w
4 y 1 y
1
D p 3 w 3 w

:KHQH[SUHVVLQJĲZYHUVXVSVHXGRVKHDUUDWHDOLQHDUDSSURDFKPD\VOLJKWO\RYHUHVWLPDWHWKH
VKHDUVWUHVVIRUVPDOOSVHXGRVKHDUUDWHYDOXHV7KLVLVVFKHPDWLFDOO\UHSUHVHQWHGLQ)LJXUH
WRJHWKHUZLWKWKHWUDQVLWLRQSRLQWZKHUHWKHUHJLPHFKDQJHVIURPODPLQDUWRWXUEXOHQWIORZ

160
140
Shear stress Ww (Pa)
120
100
80 VT
60
Water curve
40
20
0
0 100 200 300 400 500
8V/D (s-1)

Figure 5. Transition point, VT, where flow regime changes from laminar to turbulent in a
tailings slurry with Bingham like properties. The arrows indicate various
approaches for turbulent friction losses.

7KHIROORZLQJUHODWLRQVKLSLVZHOOHVWDEOLVKHGIRU%LQJKDPW\SHRIIOXLGVDQGKDVEHHQIRXQG
WRKROGIRUDSSUR[LPDWHHVWLPDWLRQVRIVT,:LOVRQHWDO

VT 222 5 y

ZKHUHĲ\LV\LHOGVWUHVVDQGȡWKHVOXUU\GHQVLW\
3.3 Free surface flow (Paper E)
)RUWKHIUHHVXUIDFHIORZLQDFKDQQHOLVFKDUDFWHUL]HGE\WKHGHSWKZHWWHGSHULPHWHUDQGIORZ
DUHD)LJXUH
Channel
y (m)
B (m)
Figure 6. Sketch of channel section with depth y and width B. The wetted perimeter, P, is
B+2y for the area, A = By.

7KH VXFFHVVLYHO\ FKDQJH LQ GHSWK DORQJ WKH IORZ LQ WKH IOXPH RYHU WKH OHQJWK ǻ/ ZDV KHUH
HYDOXDWHG LQ WHUPV RI DQ DYHUDJH VKHDU UDWH ĲZ EDVHG RQ D QHJOLJLEOH HIIHFW RI WKH YHORFLW\
KHDG

y yd gA
w u Sb
L P

ZKHUHydDQGyuDUHGRZQVWUHDPDQGXSVWUHDPVOXUU\GHSWKVRYHUWKHOHQJWKΔLDQGSbLVWKH
ERWWRP VORSH :KHQ WKH VKHDU VWUHVV KHUH LV PDWFKHG WR D SVHXGR VKHDU UDWH WKH K\GUDXOLF
GLDPHWHU'+ $3LVXVHGIRU'LQWKHSLSHOLQHH[SUHVVLRQ9'

4. Pe r f or m ance of ce ntri fug al pumps handli ng co mplex
slurries

7\SLFDO SXPS SHUIRUPDQFH HIIHFWV ZKHQ SXPSLQJ YLVFRXV 1HZWRQLDQ OLTXLGV DUH VKRZQ
)LJXUH 7KH SXPSLQJ KHDG LV GHWHUPLQHG E\ WKH SLSHOLQH FRQILJXUDWLRQ DQG WKH SLSHOLQH
UHVLVWDQFHWRIORZWKHV\VWHPFXUYH7KHRSHUDWLQJSRLQWLVGHILQHGE\WKHLQWHUVHFWLRQZLWK
WKHSXPSKHDGFXUYH

Water viscosity
Higher viscosity
H
System curve
Pump head curve
Efficiency curve
K
Flow rate
Figure 7. Schematic sketch on pump characteristics depending on fluid Newtonian viscosity.

,W IROORZV IURP )LJXUH KRZ WKH IORZ UDWH DQG KHDG LV ORZHUHG ZKHQ SXPSLQJ D YLVFRXV
1HZWRQLDQIOXLG7KHHIIHFWRIDKLJKHUYLVFRVLW\PHDQVWKDWWKHSXPSEHKDYHVDVDVPDOOHU
XQLW

7KH W\SH RI VOXUULHV FRQVLGHUHG KHUH EHKDYHV RIWHQ LQ D QRQ1HZWRQLDQ ZD\ DQG H[KLELW D
\LHOG VWUHVV :KLOH WKH GHUDWLQJ IRU D 1HZWRQLDQ IOXLG FRQYHUJHV IRU GHFUHDVLQJ IORZ UDWHV
)LJXUHDQRSSRVLWHWHQGHQF\PD\RFFXUIRUQRQ1HZWRQLDQVOXUULHVZLWKD\LHOGEHKDYLRXU
KHUHGHPRQVWUDWHGLQDQH[WUHPHVLWXDWLRQVHH)LJXUH
Slurry
H System curve
Pump head curve, water
Efficiency curve, water

K
Flow rate
Figure 8. Extreme derating effect on the pumping head for a non-Newtonian slurry with a
yield stress typical giving a large head requirements for low flow rates.

7KHV\VWHPFXUYHPDWFKHVVFKHPDWLFDOO\WKHSXPSFXUYHLQVHYHUDOSRLQWVLQ)LJXUHZKLFK
GHPRQVWUDWHV XQVWDEOH RSHUDWLRQ 7KH RSHUDWLRQ VKRZQ LV DVVRFLDWHG ZLWK WZR FRPSOLFDWLQJ
IDFWRUV FRPSDUHG WR WKH VLWXDWLRQ LQ )LJXUH IRU D YLVFRXV 1HZWRQLDQ IOXLG )LUVWO\ WKH
SRVVLEOH WHQGHQF\ RI D GLYHUJLQJ SXPS KHDG FXUYH DQG VHFRQGO\ WKH \LHOG EHKDYLRXU WKDW
JHQHUDWHDSLSHOLQHV\VWHPFXUYHWKDWLQFUHDVHYHU\OLWWOHZLWKWKHIORZUDWH7KLVPHDQVWKDW
WKHRSHUDWLQJSRLQWZLOOEHYHU\VHQVLWLYHWRWKHHIIHFWRIWKHVOXUU\RQWKHSXPSSHUIRUPDQFH

7KHUHODWLYHUHGXFWLRQRIFOHDUZDWHUKHDGDQGHIILFLHQF\IRUDFRQVWDQWIORZUDWHDQGURWDU\
VSHHGPD\EHGHILQHGE\WKHUDWLRDQGIDFWRUVVKRZQLQ)LJXUH7KHFRPSOLFDWLQJIORZUDWH
GHSHQGHQF\RQWKHSXPSSHUIRUPDQFHFXUYHVVHHQLQ)LJXUHVDQGDUHQRWLQFOXGHGLQWKH
H[SUHVVLRQV

+ +HDGUDWLR+5 ++R
+HDGUHGXFWLRQIDFWRU5+ +5
+R (IILFLHQF\UDWLR(5 KKR
K + (IILFLHQF\UHGXFWLRQIDFWRU5K (5
:DWHU
6OXUU\
KR
K
)/2:5$7(
Figure 9. Sketch defining the reduction in head and efficiency of a centrifugal pump
pumping a solids water mixture. H and H0 are heads in slurry and water service,
respectively. The corresponding efficiencies are η and η0.

7KH HQHUJ\ FRQVXPSWLRQ RI SXPSLQJ WKH VOXUU\ LQ D SLSHOLQH V\VWHP DOVR LQFOXGHV WKH WRWDO
HIILFLHQF\ T RI WKH SXPS LQ D V\VWHP $VVXPH WKDW WKH SRZHU ORVW LQ WKH PRWRU DQG LQ WKH
WUDQVPLVVLRQ FRUUHVSRQGV WR DQ HIILFLHQF\ IDFWRU M. 7KHQ T FDQ EH UHODWHG WR WKH SXPS
HIILFLHQF\ LQWKHIROORZLQJZD\

T M

7KHHIILFLHQF\UHGXFWLRQIDFWRUZDVGHILQHGDVR = 1 - / oZKHUH DQG oDUHWKHSXPS
HIILFLHQF\LQVOXUU\DQGZDWHUVHUYLFHUHVSHFWLYHO\:LWKWKHGHILQLWLRQRI R WKHQLWIROORZV
WKDW

T M O
1 R

)RU H[DPSOHZLWK R 0 DQG M LW IROORZVIURP (TWKDW TLV
DERXW7REHDEOHWRJHQHUDWHDUHTXLUHGKHDGHLQPVOXUU\WKHSXPSPXVWEHFDSDEOHRI
SURGXFLQJDKHDGRIZDWHUH0JLYHQE\(T

H0 = H/HR

7KH SXPS PXVW WKHQ RSHUDWH IRU WKH URWDU\ VSHHG FRUUHVSRQGLQJ WR H0 7KH YDOXH RI ZDWHU
HIILFLHQF\ oLVREWDLQHGIURPWKLVURWDU\VSHHG

5. Experi mental re sults
5.1 Flow resistance in pipelines and flumes (Paper B and E)
7KHIORZUHVLVWDQFHDQGUKHRORJLFDOFKDUDFWHUL]DWLRQDUHEDVHGRQYDULRXVWHVWVIURPSLSHOLQH
ORRS SXPSLQJ LQ VPDOO WR ODUJHU SLSH GLDPHWHUV DQG IRU IOXPH IORZV 3LSH VL]HV KDYH EHHQ
IURPP

2EVHUYDWLRQVRIFRQWLQXRXVIORZRYHUDOHQJWKRIDERXWPLQWKHIOXPHZLWKZLGWKDQGKHLJKW
DQG P UHVSHFWLYHO\ ZHUH FDUULHG RXW DW VORSHV IURP DERXW XS WR
$SSUR[LPDWHO\XQLIRUPFRQGLWLRQVZLWKDQHDUO\FRQVWDQWGHSWKDQGQRGHSRVLWLRQRISDUWLFOHV
ZHUH REVHUYHG 7KH FRUUHVSRQGLQJ DQDO\VLV IRU HVWLPDWLRQ RI DYHUDJH VKHDU VWUHVVHV (T LV
JLYHQLQGHWDLOVE\(QJPDQHWDO

6XPPDUL]HGSLSHDQGIOXPHZDOOVKHDUVWUHVVUHVXOWVYHUVXVSVHXGRVKHDUUDWH9'LQ)LJXUH
)RU IUHH VXUIDFH IORZ ' '+ $3 ZKHUH $ LV WKH DYHUDJH VHFWLRQ DUHD DQG 3 LV WKH
DYHUDJHZHWWHGSHULPHWHU

1000
Shear stress τw (Pa)
100 Pipeline diameter

and flume
)OXPH
P
P
10

P
P
1
0.1 1 10 100 1000
-1
8V/D (s )
Figure 10. Summary of experimentally found shear stresses versus the pseudo shear rate
8V/D for various tests in pipeline loops and in a flume.

7KHUHVXOWVLQ)LJXUH DUHEDVHGRQ &YDOXHVIURPWR ,WFDQEHVHHQWKDWWKH
\LHOG VWUHVVHV DUHLQ WKH WR 3DGRPDLQ ,QDGGLWLRQ WRWKHYDULDWLRQLQ &WKHVFDWWHU
PD\ WR VRPH H[WHQW H[SUHVV WKH YDULDELOLW\ LQ SUDFWLFH UHODWHG WR ORFDO VOXUU\ SURSHUWLHV DQG
WKLFNHQHU RSHUDWLRQ )XUWKHUPRUH WR YDULRXV WHVWLQJ FRQGLWLRQV WKDW UHSUHVHQW UHPL[LQJ DQG
PLOGO\ WR VWURQJO\ FLUFXODWLRQ WKURXJK D SXPS RU D SURGXFW WKDW KDG EHHQ VWRUHG IRU RQH WR
VHYHUDOZHHNVGHSHQGLQJRQWULDOFRQGLWLRQV

5.2 Deposition slope evaluations (Paper E)
6ORSHHYDOXDWLRQVZHUHFDUULHGRXW DWWKH6YDSSDYDDUD&RQFHQWUDWRUDQGLQDODUJHWUHQFKDW
DQRWKHUFRQFHQWUDWRU0XOWL6HUYH$%IRUDVLPLODUWDLOLQJVSURGXFW7KHGLVFKDUJHIURPWKH
SLORWVFDOHWKLFNHQLQJWHVWVZHUHZLWKPSODVWLFKRVHVDQGDW0XOWL6HUYHZLWKGLVFKDUJH
LQDWUHQFKZLWKERWWRPZLGWKRIPDQGKHLJKWDERXWP7KHGHSRVLWLRQREVHUYDWLRQVDW
WKH 6YDSSDYDDUD FRQFHQWUDWRU WRRN SODFH RQ D KRUL]RQWDO IORRU ZKLFK FUHDWHG FRQHV ZLWK
GLDPHWHUVRIWRPDQGKHLJKWVRIWRP,QDGGLWLRQEDWFKWHVWVZHUHFDUULHGRXWLQD
IOXPHZLWKZLGWKKHLJKWDQGOHQJWKRIDQGUHVSHFWLYHO\7KHYDULRXVODERUDWRU\
DQG ILHOG VHWXSV DQG GHWDLOHG UHVXOWV DUH JLYHQ E\ (QJPDQ HW DO D VXPPDU\ RI
GHSRVLWLRQVORSHVDUHVKRZQLQ)LJXUH

30
)ORRU
%DWFK
25 7UHQFK
7UHQFKFLUFXODWHG
%DWFKFLUFXODWHG
20
Slope (%)
15
10
0
38 45 50
C (%)
Figure 11. Observed deposition angles from 0.025 m diameter discharge pipelines, batch
laboratory-scale flume tests and pilot-scale trench results. Unfilled and filled
marks correspond to tests at two companies’ with similar particle size
distributions of the tailings. Solids density 2900 kg/m3.

,Q)LJXUH³FLUFXODWHG´LVFRXSOHGWRWHVWVZLWKDJLWDWHGVOXUU\WKURXJKDFLUFXODWLQJSXPS
DUUDQJHPHQW LQ WKH ORZHU SRUWLRQ RI WKH WKLFNHQHU SXPSV ZLWK PHFKDQLFDO VHDOV ,W ZDV
REVHUYHGWKDWWKLVVOXUU\IORZHGPRUHHDVLO\GRZQWKHVORSH ,W IROORZV IURP )LJXUH WKDW
DYHUDJHVORSHVRIWRZHUHREWDLQHGIURPDYROXPHWULFFRQFHQWUDWLRQRIDERXW
DQGWKDWSXPSFLUFXODWLRQLQWKHWKLFNHQHUIRUWKHWUHQFKGDWDWHQGHGWRORZHUWKHVORSHWKXV
DIIHFWLQJWKHEHKDYLRXURIWKHVOXUU\

5.3 Pump performance (Paper A and B)

$QH[DPSOHRIREVHUYHGSXPSGHUDWLQJVIRUDVLPXODWHGWDLOLQJVSURGXFWDUHJLYHQLQ)LJXUH
$GGLH 6HOOJHUQ 7KH SXPS ZDV D PRGLILHG YDQH PHWDO XQLW ZLWK DQ RSHQ
VKURXGPGLDPHWHULPSHOOHUKDYLQJDVLPSOHDXJHUOLNHLQGXFHU

16 80
14 70
12 60
Efficiency (%)
10 50
Head (m)
8 40
6 30
4 20
2 10
0 0
30 40 50 60 70 80
3
Flowrate (m /hr)
Head, Water Head, Slurry Efficiency, Water Efficiency, Slurry

Figure 12. Pump performance derating for a simulated tailings product characterized by a
yield stress of about 100 Pa. From Addie & Sellgern (2007).

,W IROORZVIURP )LJXUH WKDW WKHKHDG ZDV VWDEOHDQG +5DQG(5 ZHUHIRXQGWR EH DERXW
DQGUHVSHFWLYHO\7KHIORZUHVLVWDQFHZDVHYDOXDWHGIURPSLSHOLQHORRSUHVXOWVDQG
H[SUHVVHGLQWHUPVRIWKH \LHOGVWUHVVDQGDSODVWLFYLVFRVLW\7KHFRUUHVSRQGLQJ \LHOGVWUHVV
DQG SODVWLF YLVFRVLW\ ZHUH HVWLPDWHG WR DERXW 3D DQG 3DV UHVSHFWLYHO\ $QRWKHU
WDLOLQJVVLPXODWHG:KLWORFNDQG6HOOJUHQZLWKWKHVDPHSXPSIRUDFOD\VDQGPL[WXUH
VKRZHG+5DQG(5YDOXHVRIDQGUHVSHFWLYHO\7KHFRUUHVSRQGLQJ\LHOGVWUHVVDQG
SODVWLFYLVFRVLW\ZHUHKHUHDERXW3DDQG3DVUHVSHFWLYHO\

&RRNH SUHVHQWHG UHVXOWV IRU GLIIHUHQW WUXH WKLFNHQHG WDLOLQJV SURGXFWV ZLWK \LHOG
VWUHVVHVRIXSWR3DDQG SIURPWR3DV7KHFRUUHVSRQGLQJYDOXHVRI+5DQG
(5ZHUHLQWKHUDQJHRIWRIRUDSXPSZLWKLPSHOOHUGLDPHWHUP
5.4 Overall result summary

,W ZDV IRXQG WKDW WKH VROLGV FRQFHQWUDWLRQV E\ YROXPH & RI DERXW FKDUDFWHUL]HV WKH
SLSHOLQH SXPSLQJ KHUH HVWLPDWHG WR EH RQ WKH FRQVHUYDWLYH VLGH ZLWK UHVSHFW WR WKLFNHQHG
SURGXFWTXDOLW\UHTXLUHPHQW7KHRYHUDOOSLSHOLQHORRSUHVXOWVZHUHHYDOXDWHGIURP)LJLQ
WHUPV RI D UHSUHVHQWDWLYHZDOOVWUHVVRI3D 7KHYDULDWLRQRIVKHDUVWUHVVHVLQ)LJXUH
PD\ WR VRPH H[WHQW VKRZ WKH YDULDWLRQ LQ SUDFWLFH UHODWHG WR ORFDO VOXUU\ SURSHUWLHV DQG
WKLFNHQHURSHUDWLRQ

7KH H[SHULPHQWDO UHVXOWV FRQILUPHG WKDW WKH ORFDO FRQGLWLRQV IDYRXUHG SODFHPHQW ZLWK DQ
HOHYDWHG WKLFNHQHU ORFDWLRQ FORVH WR WKH GLVSRVDO DUHD ,Q WKLV ZD\ D QHZ KLJK SUHVVXUH
SXPSLQJ V\VWHP IRU D SUHVVXUH RI DERXW 03D LV DYRLGHG ZKLOH WKH WKLFNHQHU XQGHUIORZ
VOXUU\LVGHOLYHUHGZLWKDFHQWULIXJDOSXPSWKHVKRUWGLVWDQFHDFWXDOKHUHGXULQJWKHILUVW\HDUV
RIRSHUDWLRQ

&RQWLQXHGXVHRIFHQWULIXJDOSXPSVLQVHULHVLQVWDOODWLRQVLVFRQVLGHUHGDVWKHGLVFKDUJHSRLQW
DGYDQFHV GXULQJ WKH \HDU OLIH 7KHUHIRUH WKH XSFRPLQJ GLVFXVVLRQ LV IRFXVHG RQ WKH

SHUIRUPDQFHRI FHQWULIXJDO SXPSVLQFOXGLQJ WKH UHVXOWVRIDIROORZXSRI WKHRSHUDWLRQRI D
WKLFNHQHUXQGHUIORZSXPSLQVWDOODWLRQ

7KHSXPSSHUIRUPDQFHOLWHUDWXUHUHVXOWVDERYHIRUWUXHRUVLPXODWHGWKLFNHQHGWDLOLQJVFDQEH
VXPPDUL]HGDVPD[LPXPUHGXFWLRQVLQKHDG5+DQGHIILFLHQF\5 RIDERXWDQG
UHVSHFWLYHO\ZKHQRSHUDWLQJLQWKHEHVWHIILFLHQF\UHJLRQ

6. Discussion
6.1 Centrifugal pump performance (Paper A and B)

7KHSXPSKHDGIORZUDWHFKDUDFWHULVWLFVZKHQSXPSLQJZDWHUUHVXOWVIURPWKHVXEWUDFWLRQRI
ORVVHVIURPWKHWKHRUHWLFDOFKDUDFWHULVWLFVWUDLJKWOLQHDVVFKHPDWLFDOO\VHHQLQ)LJXUH

H ”Shock losses”
The
ore
tica
l he
”F ad
ric line
tio
nl
Actual pump os
se
H-Q curve s”
Hf
Hs
QBEP Q
Figure 13. Theoretical head curve and hydraulic losses from the ideal head curve. Partly
from Wilson et al.(2006). Hs and Hf defined in Eqs. 13 and 14, Sayers (1990).

7KH VXFFHVVLYHO\ LQFUHDVHG HIIHFW RI UHFLUFXODWLRQ DQG VKRFN ORVVHV Hs IRU GHFUHDVLQJ
WXUEXOHQWIORZUDWHVEDVLFDOO\LQFUHDVHZLWKWKHVTXDUHRIWKHIORZUDWHDW%(3PLQXVWKHDFWXDO
IORZUDWH6D\HUV

HS QBEP –Q2

7KHGLUHFWFRQGXLWIORZIULFWLRQORVVHVHfLVH[SUHVVHGDV

Hf = Q2

7KHSUREOHPZLWKIORZRIQRQ1HZWRQLDQPHGLDZLWKLQDSXPSLVWRGHILQHDUHSUHVHQWDWLYH
VKHDU UDWH 3XOOXP HW DO *UDKDP HW DO GHILQHG D FKDUDFWHULVWLF SDUDPHWHU WR
H[SUHVVWKHVKHDUUDWHZKLFKIRUWKHW\SHRIVOXUULHVFRQVLGHUHGKHUHZRXOGUHVXOWLQODPLQDU
IORZWKURXJKWKHSXPSYDQHSDVVDJHV

5KHRJUDPV ZHUH VKRZQ WRJHWKHU ZLWK YDULRXV SHUIRUPDQFH GDWD LQ :HQQEHUJ 6HOOJUHQ
7KHLUSLSHOLQHORRSUHVXOWVIRU&DERXWZHUHSUHVHQWHGLQUKHRORJLFDOWHUPVWKH
%LQJKDP\LHOGVWUHVVDQGSODVWLFYLVFRVLW\ZHUH3DDQG3DVUHVSHFWLYHO\)RXUVHWV
RI GDWD DUH UHSUHVHQWHG LQ )LJXUH LQ WHUPV RI DQ DSSDUHQW YLVFRVLW\ YHUVXV WKH VFDOLQJ
SDUDPHWHU9'

)ORZ HYDOXDWLRQV LQ WHUPV RI DSSDUHQW YLVFRVLWLHV IURP SLSHOLQHV RU RWKHU PHDVXUHPHQWV IRU
VHYHUDOLQYHVWLJDWLRQVDQGRZQWDLOLQJVVOXUU\GDWDZHUHSUHVHQWHGLQVXPPDU\UKHRJUDPDQG
WDEOHSUHVHQWDWLRQV LQ :HQQEHUJ DQG 6HOOJUHQ )RXU VHWV RI GDWD DUH UHSUHVHQWHG LQ
)LJXUH

Rotary Impeller Pump type, comments and
Notation Product C HR ER
speed diameter reference
E\PVHPLRSHQ
.DROLQ
YDQHLPSHOOHU%RRWOH
E\FORVHGYDQH
6OXGJH
LPSHOOHU6HOOJUHQHWDO
E\PRSHQVKURXGHG
&OD\
YDQHLPSHOOHU:KLWORFN
VDQG
6HOOJUHQ
Figure 14. Apparent viscosity for four sets of tests with mainly non-settling slurries with
three centrifugal pumps with various impeller diameters and configurations.

,WIROORZVIURP)LJXUHKRZWKHQRQ1HZWRQLDQEHKDYLRXULVVWURQJO\LQIOXHQFHGE\\LHOG
VWUHVVIRUGHFUHDVLQJ9'DQGWKDWDQDV\PSWRWLFYDOXHLVREWDLQHGIRUYHU\ODUJH9'

7RHYDOXDWHWKHIORZUHJLPHLQWKHSLSHOLQHVDWWDFKHGWRWKHSXPSYHORFLWLHVZHUHFDOFXODWHG
DQGFRPSDUHGWRWUDQVLWLRQYHORFLWLHVGHWHUPLQHGIURP(T

VT 22
2 5 y

7KHFRPSDULVRQVVKRZHGWKDWWKHYHORFLWLHVLQWKHVXFWLRQDQGGLVFKDUJHVHFWLRQVZHUHKLJKHU
WKDQ97LHWXUEXOHQWIRUWKHFDVHVPDUNHGDQGLQ)LJXUHDWIORZUDWHVLQWKH%(3
UHJLRQ 7KH FRUUHVSRQGLQJ VXFWLRQ IORZV PDUNHG DQG ZHUH ODPLQDU $OO VXFWLRQ VLGH
IORZVZHUHODPLQDUDW4%(3

:KHQWKHIORZHQWHUVWKHLPSHOOHUH\HLWLVH[SRVHGWRDGHJUHHFKDQJHLQGLUHFWLRQDQGD
URWDU\LQIOXHQFH7KHDFFHOHUDWLRQGHFHOHUDWLRQIORZILHOGLQWKHHQWUDQFHUHJLRQPD\DOVREH
DVVRFLDWHGZLWKLQWHUPHGLDWHODPLQDUWXUEXOHQWIORZIRU\LHOGPDWHULDO5DVKDLGD

)ULFWLRQORVVHVLQWXUEXOHQWFRQGXLWIORZGHFUHDVHLQSURSRUWLRQWRWKHGHFUHDVHLQ4UDLVHGWR
WKHVHFRQGSRZHU(T)RUQRQ1HZWRQLDQVOXUULHVIORZLQJWXUEXOHQWO\DFRQVWDQWYLVFRVLW\
RIWHQ GHILQHG E\ WKH DSSDUHQW YLVFRVLW\ IRU YHU\ ODUJH VKHDU UDWHV )LJXUH RU IRU WKH
SODVWLFYLVFRVLW\LHD1HZWRQLDQDSSURDFKIROORZLQJ:DONHUDQG*RXODV

)ULFWLRQ ORVVHV LQ ODPLQDU 1HZWRQLDQ LH FRQVWDQW YLVFRVLW\ FRQGXLW IORZ GHFUHDVH LQ
SURSRUWLRQ WR D GHFUHDVH LQ WKH IORZ UDWH )RU WKH QRQ1HZWRQLDQ \LHOG EHKDYLRU H[SUHVVHG
VFKHPDWLFDOO\LQ)LJXUHZLWKDQLQFUHDVLQJDSSDUHQWYLVFRVLW\IRUDGHFUHDVLQJIORZUDWHWKH
GHFUHDVHLQIULFWLRQORVVHVLVVPDOOW\SLFDOO\ZLWKDSRZHURILQ47KHUHIRUHZLWKDQ
DVVXPSWLRQ RI ODPLQDU FRQGLWLRQV LQ WKH SXPS D VFKHPDWLF FRPSDULVRQ ZLWK )LJ VKRZV
KRZWKHYLVFRXVIULFWLRQPD\WDNHDELJJHUVKDUHIRUORZHUIORZUDWHV7KXVTXDOLWDWLYHO\IRU
ORZHU IORZ UDWHV WKH QRQ1HZWRQLDQ EHKDYLRXU WRJHWKHU ZLWK VKRFN ORVVHV ZRXOG H[SUHVV D
ORZHULQJRIWKHKHDGFXUYHLHXQVWDEOHFRQGLWLRQV

%RRWOHLQYHVWLJDWHGWKHGHUDWLQJHIIHFWVIRUKLJK3DDQGORZ\LHOG3DVWUHVV
NDROLQVOXUULHVVHH:HQQEHUJHWDODQGIRXQGWKDWWKHYDOXHVIRU+5DERXWDQG(5
ZHUH QRW DIIHFWHG GHVSLWH WKH ELJ FKDQJH LQ \LHOG VWUHVV 7KH UHVXOWV VKRZHG WKDW WKHUH
ZHUHSUDFWLFDOO\QRGHUDWLQJRQWKHKHDG7KHYDQHVHPLRSHQPLPSHOOHUSXPSXVHG
E\%RRWOHIRUNDROLQZDVPRGLILHGZLWKDQHQODUJHGVXFWLRQLQOHWIURPWRP
DQG LQWHJUDWHG IORZ LQGXFHU VFRRSV ZKLFK SURMHFW LQWR WKH LQOHW 7KH XQH[SHFWHG QHJOLJLEOH
LQIOXHQFHRIWKH\LHOGVWUHVVZDVUHODWHGWRDGRPLQDWLRQRIWXUEXOHQWHIIHFWV;XHWDO

5HVXOWV E\ .DEDPED UHSRUWHG E\ :HQQEHUJ HW DO VKRZHG D WHQGHQF\ RI D
ORZHULQJ KHDG ZLWK D QRQ1HZWRQLDQ OLTXLG ZLWK SVHXGRSODVWLF SURSHUWLHV 7KLV LPSOLHV QR
\LHOGEHKDYLRUVHH)LJXUH7KLVLQGLFDWHVWKDWD\LHOGVWUHVVLVQRWQHFHVVDULO\DGHWHUPLQLQJ
IDFWRUIRUWKHGHYHORSPHQWRIDQXQVWDEOHKHDGFXUYH

3XPSSHUIRUPDQFHGHUDWLQJUHVXOWVIRUDQRQ1HZWRQLDQVOXUU\FDXVLQJXQVWDEOHKHDGFXUYHV
DUHVKRZQLQ)LJXUH

Figure 15. The effect of slurry on the pump head and efficiency at solids concentrations by
volume of 31 and 35 % of a scrubber sludge product. Impeller diameter 0.63 m.
y = 35 Pa and p = 0.03 Pas for 35 %. From Sellgren et al.(1999).

7KHFRPSDUDWLYHO\ORZ\LHOGVWUHVVDQGSODVWLFYLVFRVLW\IRU& LQ)LJXUHDVZHOODV
WKHWXUEXOHQWIORZSUHGLFWLRQLQWKHVXFWLRQDQGGLVFKDUJHSLSHOLQHV&DVHLQ)LJXUHGR
QRWUHIHUGLUHFWO\WRODPLQDUIORZZLWKLQWKHSXPSDQGDVVRFLDWHGKLJKIULFWLRQORVVHVIRUWKH
VKDUSGURSLQKHDG

7KH REVHUYHG VHWLQ RI WKH GHVWDELOL]DWLRQ HIIHFW PD\ EH UHODWHG WR KRZ IDU DZD\ IURP WKH
%(3UHJLRQ WKH SXPS RSHUDWHV ZLWK VXFFHVVLYHO\ PRUH UHFLUFXODWLRQ DQG VKRFN ORVVHV $Q
LQFUHDVH LQ URWDU\ VSHHG LQ WKLV GRPDLQ IRU D FRQVWDQW IORZ UDWH PD\ FRQWULEXWH WR WKH
GHVWDELOL]DWLRQ EHFDXVH WKH GLVWDQFH DZD\ IURP %(3 LQFUHDVH VHH VFKHPDWLF H[DPSOH LQ
)LJ

2
1.5 N
Relative head
B BEP- Line
N
1 A
1 2
Relative flow rate
Figure 16. An increase in head from A to B through an increase in rotary speed, N, for a
constant flow rate means operating further away from the BEP-region. From
Wennberg et al. (2008).

,QWKHVLWXDWLRQLQ)LJXUHZLWKDFRQVWDQWIORZUDWHWKHHIIHFWRIDSSDUHQWYLVFRVLW\LVQRW
FDXVHG E\ D GHFUHDVHG IORZ UDWH RQO\ WKH FKDQJH LQ ORFDO IORZ SDWWHUQ DVVRFLDWHG ZLWK WKH
LQFUHDVHLQWKHURWDU\VSHHGDQGRSHUDWLRQDZD\IURP%(3UHJLRQ7KHVFKHPDWLFVLWXDWLRQLQ
)LJXUHRFFXUUHGLQDOHVVSURQRXQFHGH[SHULPHQWDOVLWXDWLRQZKHQDWHQGHQF\RIDGURSLQ
KHDGZDVPHWZLWKDQLQFUHDVHLQURWDU\VSHHGZKHQ1LQFUHDVHGIURPWRUSPIRU
SUDFWLFDOO\FRQVWDQWYDOXHVRIIORZUDWH&DQGSLSHOLQHIULFWLRQORVVHVGXULQJDERXWPLQXWHV
:HQQEHUJHWDO

7KLV VLWXDWLRQ PD\ KDYH EHHQ FDXVHG E\ SURJUHVVLYH EXLOGXS RI PDWHULDO ZLWKLQ WKH SXPS
ZLWK JHRPHWULFDO PRGLILFDWLRQV RI IORZ SDVVDJHV ZLWKLQ WKH SXPS WKDW SURPRWH ORVVHV DQG
DIIHFWV WKH VXFWLRQ SHUIRUPDQFH 7HQGHQFLHV RI QRIORZ VWDJQDQW ]RQHV DW YDQHERXQGDULHV
ZLWKDZHOOGHILQHG%LQJKDPIOXLG8OWUH[IOXLGZLWKLQDE\PSXPSLPSHOOHU
PKDYHEHHQGRFXPHQWHGLQSLFWXUHVZLWKDQDFU\OLFVHWXS3XOOXP7KHWHQGHQFLHV
PD\ VLPXODWH EXLOWXS RI PDWHULDO ZLWKLQ WKH SXPS 7KHVH DWWDFKHG OXPSV FDQ RFFOXGH
IUDFWLRQVRIWKHSDVVDJHVZLWKLQWKHSXPSZKLFKPD\VHULRXVO\LQIOXHQFHWKHKHDGSURGXFHG

7KH DYDLODEOH DEVROXWH SUHVVXUH DW WKH SXPS VXFWLRQ VLGH VKRXOG EH ZHOODERYH WKH UHTXLUHG
DEVROXWHSUHVVXUH136+5UHODWHGWRWKHSXPSGHVLJQDQGJLYHQE\WKHSXPSPDQXIDFWXUHU
136+5LQFUHDVHVZLWKWKHIORZUDWHDQGLVQRUPDOO\FRQVLGHUHGWRKDYHWKHVDPHYDOXHDVIRU
ZDWHUZKHQSXPSLQJVOXUU\

5RXGQHYIRXQGWKDW136+5FDQEHVHYHUDOWLPHVODUJHUWKDQWKHZDWHUYDOXHEDVHGRQ
H[SHULPHQWDO UHVXOWV IURP 5XVVLDQ UHIHUHQFHV IURP WR ZLWK %LQJKDPW\SH RI
VOXUULHV IRU H[DPSOH KHDY\ PHGLD PDJQHWLWH SURGXFWV 'HWDLOHG HYDOXDWLRQV RI WKH GDWD E\
5RXGQHYVKRZHGWKDWWKHODUJHVWHIIHFWRIDQLQFUHDVHG136+5RFFXUUHGLQWKH%(3UHJLRQ
ZKHUHLWZDVDERXWWLPHVWKHZDWHUYDOXH$W4%(3WKHFRUUHVSRQGLQJYDOXHZDVRQO\
7KXV UHJXODU VXPSSXPS DUUDQJHPHQWV DUH PRUH VXVFHSWLEOH WR 136+5 WKDQ WKLFNHQHU
XQGHUIORZ LQVWDOODWLRQV DQG VXIILFLHQW VDIHW\ IDFWRUV VKRXOG EH XVHG +RZHYHU LI QR GLUHFW
QHJDWLYHHIIHFWLVQRWHGDW%(3IORZVWKHQDQHIIHFWDW4%(3LVOHVVOLNHO\

5RXGQHY H[SUHVV WKDW WKH JUHDWHU 136+5 UHTXLUHPHQW FDQ EH UHODWHG WR EORFNDJH
WHQGHQFLHVLQWKHHQWUDQFHUHJLRQE\WKH\LHOGEHKDYLRUUHVXOWLQJLQKLJKHUYHORFLWLHVDQGWKXV
ORZHU SUHVVXUH $GYHUVH SXPS SHUIRUPDQFH HIIHFWV UHODWHG WR WKH \LHOG EHKDYLRXU ZDV DOVR
SRLQWHGRXWE\5DVKDLGDZLWKIRUPDWLRQRIVWDJQDQW³FDYHUQV´LQURWDWLRQGLVNV\VWHPV
6WURQJYLVFRXVHIIHFWVRQWKHIORZSDWWHUQQHDUWKHLPSHOOHULQOHWZDVREVHUYHGE\/L

7KH GLVFXVVHG PHFKDQLVPV KDYH VRPH VLPLODULWLHV WR SRVVLEOH EORFNDJH IURP YDSRXU LQ WKH
HQWUDQFH UHJLRQ VRPHWLPHV REVHUYHG IRU OLTXLG SXPSLQJ DW ORZ IORZ UDWHV ZLWK VWURQJ ORFDO
UHFLUFXODWLRQFDXVLQJFDYLWDWLRQDOFRQGLWLRQV*LUGKDUDQG0RQL])LJXUH

Figure 17. The pressure on the suction side of the pump and the eye of the impeller when the
suction pressure is lowered bubbles starts forming in the eye of the impeller.
From Girdhar and Moniz (2004).

)ROORZLQJWKHGLVFXVVLRQDERYHDWKLFNHQHUXQGHUIORZSXPSVKRXOGQRWEHH[SRVHGGLUHFWO\WR
D 136+5 SUREOHP EHFDXVH LW LV QRUPDOO\ ORFDWHG D FRQVLGHUDEOH GHSWK EHORZ WKH VOXUU\
VXUIDFH OHYHO 7KHUHIRUH WKH DEVROXWH SUHVVXUH UHTXLUHPHQW LV QRUPDOO\ IXOILOOHG LQ WKH
GHWHUPLQLQJ%(3UHJLRQ

$QLQVWDOODWLRQZDVLGHQWLILHGZKHUHDFHQWULIXJDOSXPSRSHUDWHGXQGHUDWKLFNHQHUZLWKKLJK
GHQVLWLHVIRUDSDVWHXVHGIRUPLQHEDFNILOOLQJDWWKH1HZ%ROLGHQ*DUSHQEHUJRSHUDWLRQV
(ULNVVRQ7KHWKLFNHQHGDQGIORFFXODWHGVOXUU\IORZVE\JUDYLW\DERXWPWRWKHEDVH
RIWKHWKLFNHQHUWRWKHSXPSIURPZKHUHLWLVWUDQVSRUWHGDERXWP6DPSOHVZHUHWDNHQRXW
IURPUKHRORJLFDOPHDVXUHPHQWVDQGIORZUDWHDQGSUHVVXUHPHDVXUHPHQWVZHUHFDUULHGRXW
5KHRJUDPUHVXOWVDUHVKRZQLQ)LJXUHDQGFRPSDUHGWRWKHSLSHOLQHORRSUHVXOWVLQ)LJXUH

Figure 19. Rheogram for C-values from 43.2-46.5 % for the backfilling paste measured with
a vane-type Bohlin-viscometer. Comparison with the pipeline-loop results in
Figure 10.

7KHSDUWLFOHVL]HGLVWULEXWLRQIURPWKHWKLFNHQHUXQGHUIORZKDGDQDYHUDJHVL]HRIGDERXW
PZLWKRIWKHSDUWLFOHVILQHUWKDQ P7KHWRSVL]HZDVDERXW PDQGWKH
VROLGV GHQVLW\ ZDV DERXW NJP (YDOXDWLRQ RI WKH IULFWLRQV ORVVHV LQ WKH SLSH V\VWHP
IURPIORZUDWHDQGSUHVVXUHPHDVXUHPHQWVVKRZHGWKDWWKHDYHUDJHZDOOVKHDUVWUHVV ZZDV
DERXW 3D 7KH YDQHYLVFRPHWHU UHVXOWV LQ )LJXUH LQGLFDWHV VLPLODU YDOXH IRU DYHUDJH
RSHUDWLQJFRQGLWLRQV

7KHSXPSZDVUXQQLQJZLWKURWDWLRQDOVSHHGVRIUSPDQGWKHRSHUDWLQJGRPDLQLV
VFKHPDWLFDOO\VKRZQLQ)LJXUH

40
180
0r
pm
35
160
30 0 rp
m
BEP-Line
Operating
25 domain
140
H (m) 0 rp
m
20
15
10
50 100 150 200 250 300
Q (m3/h)
Figure 20. Schematic working domain for the pump when handling high concentrations
shown in Fig.19. The pump was a 3-vane all-metal unit with a semi-open impeller
with diameter 0.35 m.

,W IROORZV VFKHPDWLFDOO\ IURP )LJXUH KRZ WKH SXPS LV ZRUNLQJ IDU D ZD\ IURP WKH EHVW
HIILFLHQF\OLQHDQGDWDERXW4%(3$Q\RSHUDWLQJSUREOHPVZLWKDQXQVWDEOHKHDGFXUYHKDV
QRWEHHQH[SHULHQFHG
6.2 Depositional conditions (Paper C and E)

&HQWULIXJDO SXPSV DUH NQRZQ WR LPSDUW KLJK VKHDU LQWHQVLWLHV UHVXOWLQJ LQ D UHGXFHG VOXUU\
\LHOG VWUHVV )XUWKHUPRUHWKHVOXUU\LV H[SRVHGWR DJLQJDQGYDULRXVGHJUHHV RIVKHDULQ WKH
GRZQVWUHDP SLSHOLQH GLVWULEXWLRQ V\VWHP 3XEOLVKHG UHVXOWV DUH GLIILFXOW WR HYDOXDWH DQG
VRPHWLPHVFRQWUDGLFWRU\DVGLVFXVVHGE\:HQQEHUJDQG6HOOJUHQ

3DUWV RI WKH XQGHUIORZ LQ SDVWH WKLFNHQHUV DUH WRGD\ RIWHQ FLUFXODWHG EDFN LQWR WKH EDVH
WKURXJK SXPSLQJ LQ RUGHU WR DJLWDWH WKH SDVWH 3LORW VFDOH WKLFNHQLQJ DQG WUHQFK GHSRVLWLRQ
WHVWV IRU D VLPLODU WDLOLQJV SURGXFW DV LQYHVWLJDWHG KHUH ZLWK DQG ZLWKRXW FLUFXODWLRQ ZLWK D
FHQWULIXJDOSXPSVKRZHGWKDWWKHDJLWDWHGSDVWHIORZHGPRUHHDVLO\GRZQWKHVORSHUHVXOWLQJ
LQDORZHUGHSRVLWLRQDQJOH

7KHUHIRUH LQGHSHQGHQW RI WKH V\VWHP FRQILJXUDWLRQ VORSH IRUPLQJ PHFKDQLVPV FDQ EH
FRQVLGHUHG WR EH UHODWHG WR VKHDU VWUHVVHV WKDW DUH PXFK ORZHU WKDQ WKRVH REVHUYHG LQ WKH
WKLFNHQHUXQGHUIORZ

5HFHQWDSSURDFKHVIRUVORSHHYDOXDWLRQVIRUH[DPSOHE\3LURX]HWDODQGRWKHUVPD\
SURYLGH D PHDQV RI SUHGLFWLQJ WKH EHDFK VORSH DQJOH ZLWK WDLOLQJV VOXUULHV RI DUELWUDU\
UKHRORJLFDOFKDUDFWHULVDWLRQ3LURX]HWDODSSOLHGYDULRXVQRQ1HZWRQLDQSLSHOLQHIULFWLRQORVV
PRGHOVWRVLPXODWHIUHHVXUIDFHIORZORVVHVIRUDWDLOLQJVVOXUU\7KHUHSRUWHGILQGLQJVH[SODLQ
WR VRPH H[WHQW WKH GLIILFXOW\ ZLWK ODERUDWRU\ VFDOH IOXPH WHVWV EHFDXVH RI WKH SUREOHP WR
REWDLQWXUEXOHQWIORZFRQGLWLRQV

$Q DOWHUQDWLYH WR D UKHRORJLFDOO\EDVHG VORSH PRGHOOLQJ DSSURDFK ZDV FRQVLGHUHG ZLWK
UHIHUHQFHVWR'H*URRWHWDODQG&85ZKRLQYHVWLJDWHGWKHVORSHUHTXLUHPHQW
IRUDEDODQFHEHWZHHQVHGLPHQWDWLRQDQGHURVLRQDORQJDEHDFKIRUPHGE\GUHGJLQJPDWHULDO
IURP SDUDOOHO GLVFKDUJHV IURP VHYHUDO ODUJH SLSHOLQHV 7KH\ LQYHVWLJDWHG VORSH DQJOHV IRU
YDULRXVQDUURZO\JUDGHGILQHVDQGPHGLXPFRQFHQWUDWLRQVOXUULHVZLWKSDUWLFOHVGRZQWRDERXW
PLHWZRFRPSRQHQW³VHWWOLQJ´W\SHRIPL[WXUHV

$GDSWLRQRIWKHLUDSSURDFKLQDVLPSOLILHGZD\IRUWKHDFWXDOWDLOLQJVH[SHULPHQWVH[SUHVVHV
KRZ VPDOO SLSHOLQH GLVFKDUJH UDWHV WHQG WR JLYH KLJKHU VORSHV 7KH UHVXOW LQGLFDWHV WKDW
GLVSRVDO VORSHV IRU WKH DSSOLFDWLRQV FRQVLGHUHG KHUH VKRXOG EH DERXW WR EHFDXVH
REVHUYHG VORSHV GXULQJ WKH WHVWV ZHUH WR GHSHQGHQW RQ WKH FLUFXODWLQJ SXPS ZDV
RSHUDWLQJRUQRWLQWKHSLORWVFDOHWKLFNHQHU

$ FHQWUDO YDULDEOH LQ WKH DSSURDFK DERYH LV D FKDUDFWHULVWLF ZLGWK RI WKH VOXUU\ IORZ RQ
GHSRVLWHGWDLOLQJVLQWKHYLFLQLW\RIWKHGLVFKDUJLQJSLSHOLQHOHDGLQJXSWRWKHGHILQLWLRQRID
VSHFLILFIORZUDWHLQPV:LQWHUZHUSHWDO)XUWKHUDQDO\VLVRIWKHLUGDWDLQVPDOODQG
ODUJH IOXPHV DQG SKRWR FRPSDULVRQV KHUH LQFOXGLQJ ILHOG RSHUDWLRQV FRQILUP WKH WR
HVWLPDWLRQVDERYH

7KH IORZ UDWH GHSHQGHQFH RQ WKH VORSH PHDQV IOH[LELOLW\ EHFDXVH D VSOLW RI WKH GLVFKDUJH
SLSHOLQHIORZLQWRWZRVWUHDPVJLYHKLJKHUVORSHV

7. CONCLUSIONS

:LWKDQHOHYDWHGWKLFNHQHUORFDWLRQFORVHWRWKHGLVSRVDODUHDDQHZKLJKSUHVVXUHSXPSLQJ
V\VWHP IRU D SUHVVXUH RI DERXW 03D LV DYRLGHG ZKLOH WKH WKLFNHQHU XQGHUIORZ VOXUU\ LV
GHOLYHUHGZLWKDFHQWULIXJDOSXPSDVKRUWGLVWDQFHGXULQJWKHILUVW\HDUVRIRSHUDWLRQ7RPHHW
WKHPRVWHIIHFWLYHORQJWHUPVROXWLRQIRUSLSHOLQHGLVWULEXWLRQDQGSODFHPHQWFRQWLQXHGXVHRI
FHQWULIXJDOSXPSVLQVHULHVLQVWDOODWLRQVLVFRQVLGHUHGDVWKHGLVFKDUJHSRLQWDGYDQFHVGXULQJ
WKH\HDUOLIH

$VROLGVFRQFHQWUDWLRQVE\YROXPH&RIDERXWFKDUDFWHUL]HVWKHSLSHOLQHSXPSLQJKHUH
HVWLPDWHGWREHRQWKHFRQVHUYDWLYHVLGHZLWKUHVSHFWWRWKLFNHQHGSURGXFWTXDOLW\
UHTXLUHPHQW7KHRYHUDOOSLSHOLQHORRSUHVXOWVZHUHHYDOXDWHGLQWHUPVRIDUHSUHVHQWDWLYHSLSH
ZDOOVWUHVVRI3DIRU&DERXW,QUKHRORJLFDOWHUPVWKH%LQJKDP\LHOGVWUHVVDQG
SODVWLFYLVFRVLW\ZHUH3DDQG3DVUHVSHFWLYHO\

7KHSHUIRUPDQFHRIDFHQWULIXJDOSXPSLVDIIHFWHGZKHQKDQGOLQJVOXUULHVRUYLVFRXV
IOXLGVZKLFKVKRZVXSDVUHGXFWLRQVRIWKHFOHDUZDWHUKHDGDQGHIILFLHQF\FXUYHV7KH
PD[LPXP UHGXFWLRQV LQ KHDG DQG HIILFLHQF\ ZHUH DERXW DQG UHVSHFWLYHO\
ZKHQ RSHUDWLQJ DW WKH GHVLJQ IORZ UDWH 4%(3 LQ WKH EHVW HIILFLHQF\ UHJLRQ 3XPS
LPSHOOHUVL]HVOHVVWKDQDERXWP

$VRSSRVHGWR1HZWRQLDQIOXLGVDGLYHUJLQJWHQGHQF\IURPWKHZDWHUKHDGFXUYHIRU
GHFUHDVLQJ IORZ UDWHV KDV VRPHWLPHV EHHQ UHSRUWHG EHORZ DERXW 4%(3 IRU VOXUULHV
WKDW VKRZ D KLJKO\ QRQ1HZWRQLDQ EHKDYLRXU 7KH GLYHUJLQJ WHQGHQF\ PD\ DOVR EH
UHODWHGWRWKHVXFWLRQSHUIRUPDQFH

7KHVXFWLRQVLGHSUHVVXUHUHTXLUHPHQWLVQRUPDOO\IXOILOOHGZLWKDODUJHPDUJLQ IRUD
WKLFNHQHU XQGHUIORZ SXPS ,W ZDV REVHUYHG LQ DQ LQVWDOODWLRQ KRZ D VWDQGDUG SXPS
RSHUDWLQJ DW 4%(3 XQGHU D WKLFNHQHU VKRZHG QR XQVWDEOH KHDG WHQGHQFLHV IRU D
WDLOLQJVVOXUU\ZLWKZDOOVWUHVVHVRIXSWRDERXW3D

&HQWULIXJDO SXPSV DUH NQRZQ WR LPSDUW KLJK VKHDU LQWHQVLWLHV UHVXOWLQJ LQ D UHGXFHG
VOXUU\\LHOGVWUHVV$QXQGHUIORZSXPSLVWKXVH[SRVHGWRWKHPD[LPXP\LHOGHIIHFWV
ZKLOHGRZQVWUHDP SXPSVDUHDIIHFWHGOHVV7KHUHIRUHGRZQVWUHDP SXPSV VKRXOGEH
DUUDQJHGZLWKVXIILFLHQWPDUJLQVIRUDQLQFUHDVHGVXFWLRQVLGHSUHVVXUHUHTXLUHPHQW

7HQGHQFLHV RI QRIORZ VWDJQDQW ]RQHV DW YDQHERXQGDULHV ZLWKLQ D ZHOOGHILQHG
%LQJKDPIOXLGKDYHEHHQUHSRUWHG:LWKDGRPLQDWLRQRIODPLQDUIORZLQWKHYLFLQLW\
RIVXUIDFHVERWKVWDJQDQW]RQHVDQGKLJKIULFWLRQORVVHVFDQEHVWURQJO\UHODWHGWRWKH
\LHOGEHKDYLRXUDQGWRWKHVXFWLRQSHUIRUPDQFH

6ORSHV RI GHSRVLWHG WDLOLQJV KDYH KHUH EHHQ HVWLPDWHG WR DERXW IRU VOXUU\
SURSHUWLHV WKDW JLYHV D FRQFHSWXDOO\ HYHQ VORSH ZLWK QR VHJUHJDWLRQ RI SDUWLFOHV DQG
YLUWXDOO\QRGUDLQDJHRIZDWHU

8. Future work

7KHSXPSSHUIRUPDQFHGLVFXVVLRQDERYHGRHVQRWLQFOXGHDQ\GHWDLOHGDQDO\VHVRIWKHYDULRXV
SXPS FRQILJXUDWLRQV DQG GHVLJQV WKH LQWHQWLRQ LV PHUHO\ WR IRUP D EDFNJURXQG IRU IXUWKHU
GLVFXVVLRQV ZLWK LQWHUHVWHG SDUWLHV 7KH DYDLODELOLW\ RI SXPS SHUIRUPDQFH GDWD IRU WUXH
WKLFNHQHGWDLOLQJVLVH[SHFWHGWRLQFUHDVHLQWKHQHDUIXWXUH7KHUHPD\WKXVEHDQRSSRUWXQLW\
WR TXDQWLI\ DQG YDOLGDWH WKH SXPS SHUIRUPDQFH GLVFXVVLRQ LQ WKLV VWXG\ 'HVSLWH WKDW
FRQVLGHUDEO\DGYDQFHVKDYHEHHQPDGHUHFHQWO\LQXQGHUVWDQGLQJWKHPHFKDQLVPRIIRUPDWLRQ
RI WKH VORSHV RI GHSRVLWHG WDLOLQJV WKLV LV DQ DUHD ZHUH PRUH ZRUN QHHG WR EH XQGHUWDNHQ
-HZHOO

9. REFERENCES

$GGLH*5DQG6HOOJUHQ$$ILUVWVOXUU\SXPSVWDQGDUGDQGVRPHLPSOLFDWLRQVIRU
SDVWHDSSOLFDWLRQV3URFHHGLQJVRIWKH7HQWK,QWHUQDWLRQDO6HPLQDURQ3DVWDQG7KLFNHQHG
7DLOLQJV0DUFK3HUWK$XVWUDOLD

%RRWOH0-3UDFWLFDODVSHFWVRI7UDQVSRUWLQJ3DVWZLWK5RWRG\QDPLF6OXUU\3XPSV
3URFHHGLQJVRIWKH1LQWK,QWHUQDWLRQDO6HPLQDURQ3DVWDQG7KLFNHQHG7DLOLQJV$SULO
/LPHULFN,UHODQGSS

&RRNH57KLFNHQHGDQGSDVWHWDLOLQJVSLSHOLQHV\VWHPV'HVLJQSURFHGXUHSDUW
3URFHHGLQJVRIWKH7HQWK,QWHUQDWLRQDO6HPLQDURQ3DVWDQG7KLFNHQHG7DLOLQJV0DUFK
3HUWK$XVWUDOLDSS

&85$UWLILFLDOVDQGILOOVLQZDWHU&HQWUHIRUFLYLOHQJUHVHDUFKDQGFRGHV7KH
1HWKHUODQGV$$%DONHPD5RWWHUGDP

'H*URRW0%+HH]HQ)70DVWEHUJHQ'5DQG6WHIHVV+6ORSHVDQGGHQVLWLHVRI
K\GUDXOLFDOO\SODFHGVDQGV+\GUDXOLF)LOOVWUXFWXUHV$6&(*HRWHFKQLFDO'LYLVLRQ6SHFLDOLW\
&RQIHUHQFH&RORUDGR6WDWH8QLYHUVLW\)W&ROOLQV86$

(QJPDQ0$6HOOJUHQc6XQGTYLVW,*ROGNXKODQG7:HQQEHUJ8VHUVSHUVSHFWLYH
RQWKHGHVLJQRIKLJKGHQVLW\EDVHPHWDOWDLOLQJVV\VWHPV3URFHHGLQJVWK,QW
&RQIHUHQFHRQ7DLOLQJVDQG0LQH:DVWH¶9DLO&286$2FWREHU

(ULNVVRQ23HUVRQDOFRPPXQLFDWLRQ0HWVR0LQHUDOV6ZHGHQ$%

*LUGKDU3DQG0RQL]23UDFWLFDOFHQWULIXJDOSXPSV'HVLJQRSHUDWLRQDQG
PDLQWHQDQFH(OVHYLHU

*UDKDP/-:3XOOXP/6ODWWHU36HU\*5XGPDQ0&HQWULIXJDOSXPS
SHUIRUPDQFHFDOFXODWLRQIRUKRPRJHQHRXVDQGFRPSOH[KHWHURJHQHRXVVXVSHQVLRQV7KH
&DQDGLDQ-RI&KHPLFDO(QJSS

+ROORZ-7/LQ+.:DOVK'(DQG:KLWH'0'HYHORSPHQWRIDQHQHUJ\EDODQFH
PRGHOWRSUHGLFWWKHHFRQRPLFLPSDFWRILQVWDOOLQJDWDLOLQJVZDVKWKLFNHQHUDWWKH)RUW.QR[
0LQH)DLUEDQNV$ODVND0LQHUDOVDQG0HWDOOXUJLFDO3URFHVVLQJ9RO1R0D\SS

-HZHOO5-(QVXULQJWKHFUHGLELOLW\RIWKLFNHQHGWHFKQRORJ\3URFHHGLQJVRIWKHWK
,QWHUQDWLRQDOVHPLQDURQ3DVWHDQG7KLFNHQLQJ7DLOLQJV0D\7RURQWR&DQDGDSS

.DEDPED%0(YDOXDWLRQRIGHUDWLQJSURFHGXUHVIRUFHQWULIXJDOSXPSSHUIRUPDQFH
0DJLVWHU7HFKQRORJLDHWKHVLV&DSH3HQLQVXOD8QLYHUVLW\RI7HFQRORJ\6RXWK$IULFD

/L:*(IIHFWVRIYLVFRVLW\RQIOXLGVRQ&HQWULIXJDOSXPSSHUIRUPDQFHDQGIORZ
SDWWHUQLQWKHLPSHOOHU,QWHUQDWLRQDO-RXUQDORI+HDWDQG)OXLG)ORZSS

0RRQH\0--5KHRORJ\

2[HQIRUG-DQG/RUG(5&DQDGLDQH[SHULHQFHLQWKHDSSOLFDWLRQRISDVWHDQG
WKLFNHQHGWDLOLQJVIRUVXUIDFHGLVSRVDO3URFHHGLQJVRIWKHWK,QWHUQDWLRQDO6HPLQDURQ3DVWH
DQG7KLFNHQHG7DLOLQJV$SULO/LPHULFN,UHODQGSS

3LURX]%.DQG:LOOLDPV33UHGLFWLRQRIQRQVHJUHJDWLQJWKLFNHQLQJWDLOLQJVEHDFK
VORSHVDQHZPHWKRG3URFHHGLQJVRIWKHWK,QWHUQDWLRQDO6HPLQDURQ3DVWHDQG7KLFNHQHG
7DLOLQJV3HUWK$XVWUDOLD0DUFKSS

3XOOXP/3HUVRQDOFRPPXQLFDWLRQFRXUWHV\RIWKH7KHUPDODQG)OXLGV(QJLQHHULQJ
*URXS&6,52$XVWUDOLD

3XOOXP/*UDKDP/-:5XGPDQ0$'H5DWLQJ0HWKRGIRU&HQWULIXJDO3XPSV
3XPSLQJ³3DVWH´3URFHHGLQJVRIWKH7HQWK,QWHUQDWLRQDO6HPLQDURQ3DVWDQG7KLFNHQHG
7DLOLQJV0DUFK3HUWK$XVWUDOLDSS

3XOOXP/3LSHOLQLQJWDLOLQJVSDVWHVDQGEDFNILOO3UHGLFWLRQRIQRQVHJUHJDWLQJ
WKLFNHQLQJWDLOLQJVEHDFKVORSHVDQHZPHWKRG3URFHHGLQJVRIWKHWK,QWHUQDWLRQDO6HPLQDU
RQ3DVWHDQG7KLFNHQHG7DLOLQJV3HUWK$XVWUDOLD0DUFKSS

5DELQRZLWVFK%=3K\VLN&KHP6HU$

5DVKDLGD$$)ORZRIDQRQ1HZWRQLDQ%LQJKDPSODVWLFIOXLGRYHUDURWDWLQJGLVN
'LVVHUWDWLRQ'HSDUWPHQWRI0HFKDQLFDO(QJLQHHULQJ8QLYHUVLW\RI6DVNDWFKHZDQ6DVNDWRRQ
6DVNDWFKHZDQ, 8QLYHUVLW\RI6DVNDWFKHZDQ6DVNDWRRQ6DVNDWFKHZDQ&DQDGD61$

5RELQVN\(,7KLFNHQHG7DLOLQJV'LVSRVDOLQWKH0LQLQJ,QGXVWU\4XHEHFRU
3ULQWSDN7RURQWR&DQDGD

5RXGQHY$66OXUU\SXPSVXFWLRQSHUIRUPDQFHFRQVLGHUDWLRQV3URFWK,QW
&RQIHUHQFHRQ+\GURWUDQVSRUW%+5*URXS6DQWLDJR&KLOHSS

6D\HUV$7+\GUDXOLFDQGFRPSUHVVLEOHIORZWXUERPDFKLQHV0F*5$:+,//

6HOOJUHQ$$GGLH* <X:&(IIHFWVRIQRQ1HZWRQLDQPLQHUDOVXVSHQVLRQVRQ
WKHSHUIRUPDQFHRIFHQWULIXJDOSXPSV-RXUQDORI0LQHUDO3URFHVVLQJDQG([WUDFWLRQ
0HWDOOXUJ\5HYLHZ9ROSS

6KRRN&$5RFR0&6OXUU\IORZSULQFLSOHVDQGSUDFWLFH%XWWHUZRUWK+HLQHPDQQ
86$

:DONHU &, *RXODV $ 3HUIRUPDQFH FKDUDFWHULVWLFV RI FHQWULIXJDO SXPSV ZKHQ
KDQGOLQJQRQ1HZWRQLDQKRPRJHQHRXVVOXUULHV3URF,QVWQ0HFK(QJUV9RO$1R
SS

:HQQEHUJ76HOOJUHQ$6XQGTYLVWc)HDVLELOLW\FRQVLGHUDWLRQVRIKLJKGHQVLW\
GLVSRVDODWWKH/.$%6YDSSDYDDUDFRQFHQWUDWRU3URFHHGLQJV7KH,QW&RQIHUHQFH6HFXULQJ
WKH)XWXUH-XQH6NHOOHIWHn6ZHGHQ

:HQQEHUJ7DQG6HOOJUHQ$3XPSLQJHYDOXDWLRQVZLWKSDVWHWDLOLQJVWKLFNHQHGFORVH
WRWKHVXUIDFHGLVSRVDODUHD3URFHHGLQJV7KHWK,QW&RQIRQWKH+\GUDXOLF7UDQVSRUWRI
6ROLGV&DSHWRZQ6RXWK$IULFD0D\SS

:HQQEHUJ76HOOJUHQ$:KLWORFN/3UHGLFWLQJWKHSHUIRUPDQFHRIFHQWULIXJDO
SXPSVZKHQKDQGOLQJFRPSOH[VOXUULHV3URFHHGLQJVWK,QWHUQDWLRQDO&RQIHUHQFHRQ
7UDQVSRUW 6HGLPHQWDWLRQRI6ROLGVSDUWLFOHV-XQH6DLQW3HWHUVEXUJ5XVVLDSS

:HQQEHUJ76HOOJUHQ$*ROGNXKO,5KHRORJLFDODQG'HSRVLWLRQDOFKDUDFWHULVDWLRQ
RISDVWHOLNHWDLOLQJVVOXUULHV3URFHHGLQJV&RQIHUHQFHLQ0LQHUDOV(QJLQHHULQJ)HEUXDU\
/XOHn6ZHGHQ

:LOVRQ . & $GGLH *5 6HOOJUHQ $ DQG &OLIW 5 6OXUU\ WUDQVSRUW XVLQJ FHQWULIXJDO
SXPSVUG(G%ODFNLH$FDGHPLFDQG3URIHVVLRQDO&KDSPDQDQG+DOO/RQGRQ8.

:LOVRQ.&DQG6HOOJUHQ$(IIHFWRIVWURQJO\QRQ1HZWRQLDQVOXUULHVRQWKHKHDGDQG
HIILFLHQF\RIFHQWULIXJDOSXPSV, 3URFHHGLQJVWK,QWHUQDWLRQDOFRQIHUHQFHRQWUDQVSRUW
VHGLPHQWDWLRQRIVROLGSDUWLFOHV*HRUJLD6HSWHPEHU

:LQWHUZHUS-&GH*URRW0%0DVWEHUJHQ'5DQG9HUZRHUW+
+\SHUFRQFHQWUDWHGVDQGZDWHUPL[WXUHIORZVRYHUDIODWEHG-+\GU(QJ9ROSS

;X-7LSPDQ5*LOOLHV56KRRN&$&HQWULIXJDOSXPSSHUIRUPDQFHZLWK1HZWRQLDQ
DQGQRQ1HZWRQLDQVOXUULHV3URFHHGLQJV7KHWK,QW&RQIRQWKH+\GUDXOLF7UDQVSRUWRI
6ROLGV%DQII&DQDGD%+5*URXS8.0D\SS

10. NOTATIONS

Nomenclature

$ $YHUDJHVHFWLRQDUHD P
% :LGWK P
%(3 %HVWHIILFLHQF\SRLQW
& 9ROXPHWULFFRQFHQWUDWLRQRIVROLGV
&Z 6ROLGVFRQFHQWUDWLRQE\PDVV
' 'LDPHWHU P
'+ +\GUDXOLFGLDPHWHU P
G $YHUDJHSDUWLFOHVL]H ȝP
(5 (IILFLHQF\UHGXFWLRQ
J *UDYLWDWLRQDODFFHOHUDWLRQ PV
+ 7RWDOSXPSLQJKHDG P
+I &RQGXLWIORZIULFWLRQORVV P
+R +HDGLQPHWHUVRIZDWHU P
+5 +HDGUHGXFWLRQ
+V 6KRFNORVVHV P
M )ULFWLRQJUDGLHQWLQKHLJKWRIPL[WXUHOHQJWKRISLSH PP
/ &KDQQHOVHJPHQWOHQJWK P
0 0DVVIORZUDWHRIZDWHU WRQQHVK
0V 0DVVIORZUDWHRIGU\VROLGV WRQQHVK
1 5RWDU\VSHHG UHYPLQ
n 6ORSH(T
136+5 1HWSRVLWLYHVXFWLRQKHDGUHTXLUHG P
S [ 3UHVVXUHJUDGLHQW 3DP
4 9ROXPHWULFIORZUDWH PV
5+ +HDGUHGXFWLRQIDFWRU
5Ș (IILFLHQF\UHGXFWLRQIDFWRU
9 0HDQYHORFLW\ PV
97 9DOXHRIPHDQYHORFLW\DWWUDQVLWLRQSRLQW PV
6E 6ORSHRIRSHQFKDQQHO
6V 5HODWLYHGHQVLW\RIVROLGV
4%(3 9ROXPHWULFIORZDWEHVWHIILFLHQF\SRLQW PV
\ )ORZGHSWK P
\G 'RZQVWUHDPGHSWK P
\X 8SVWUHDPGHSWK P
γ 6KHDUUDWH V
η 3XPSHIILFLHQF\
ηM 0RWRUDQGWUDQVPLVVLRQHIILFLHQF\
η0 3XPSZDWHUHIILFLHQF\
ηT 7RWDOHIILFLHQF\
μ 1HZWRQLDQYLVFRVLW\ 3DV
a $SSDUHQWYLVFRVLW\ 3DV
P 3ODVWLFYLVFRVLW\ 3DV
μ $SSDUHQWYLVFRVLW\DWODUJHVKHDUUDWHV 3DV
ρ 'HQVLW\RIVOXUU\ NJP

ρs 'HQVLW\RIVROLGV NJP
ρ0 'HQVLW\RIZDWHU NJP
τ 6KHDUVWUHVV 3D
τy <LHOGVKHDUVWUHVV 3D
τw 3LSHZDOOVKHDUVWUHVV 3D

Paper A
Predicting the performance of centrifugal pumps when handling complex slurries
Wennberg T, Sellgren A and Whitlock L
Proceedings, 14th International Conference on Transport & Sedimentation of Solids particles,

June 23-25 2008, Saint Petersburg, Russia
PREDICTING THE PERFORMANCE OF CENTRIFUGAL PUMPS
WHEN HANDLING COMPLEX SLURRIES
Thord Wennberg1, Anders Sellgren2, Lee Whitlock3

1
Luossavaara-Kiirunavaara AB, Sweden,Thord.Wennberg@lkab.com
2
Luleå University of Technology ,Sweden ,asg@ltu.se
3
GIW Industries Inc., U.S.A.,LWhitlock@giwindustries.com
Abstract: The term complex here refers to the intermediate area between homogeneously and
heterogeneously flowing slurries. Typical examples are residual products (“tailings”) from the
mining industry with average particle sizes of 20 to 100 microns. The performance derating are of
special interest when centrifugal slurry pumps are applied to these slurries at high solids
concentrations. The Hydraulic Institute’s slurry (ANSI/HI 2005) and viscous Newtonian liquid
(ANSI/HI 2004) standards are briefly covered with focus on applicable viscosities for non-
Newtonian slurry behaviour. Various experimental results are discussed and found to give
maximum deratings in head and efficiency of about 10 and 15 %, respectively, provided stable
head curves can be maintained. Magnified shock losses, circulatory flows and blockage of
slurry and possibly vapour tendencies in the pump entrance region may be considered as
mechanisms behind unstable head curves together with the flow behaviour of the slurry.
KEY WORDS: complex slurries, pump performance derating, entrance losses
1. INTRODUCTION
The term complex here refers to the intermediate area between homogeneously and
heterogeneously flowing slurries or ”non-settling” and “settling” types of mixtures,
respectively. Typical examples are residual products (“tailings”) from the mining
industry with normal average particle sizes of 20 to 100 microns, see for example
Wennberg and Sellgren (2007). Today there is an increased interest in using less water,
which means handling of tailings slurries at high solids concentrations. Criteria for
thickening and pumpability are often referred to in rheological terms, for example yield
stress, i.e. homogeneous flow. It is known that these slurries may not behave in a
completely non-settling way flowing homogeneously during pipeline transportation
because larger particles may show settling tendencies.
The performance of centrifugal pumps is influenced by the solids in the slurry

through deratings of the clear water head and efficiency. The Hydraulic Institute has
developed a new standard (ANSI/HI 2005) for centrifugal slurry pumps were guidelines
are given to predict the performance when handling typical settling or non-settling
slurries.
The performance of a centrifugal pump is affected when handling slurries. The
relative reduction of the clear water head and efficiency for a constant flow rate and
rotary speed may be defined by the ratios and factors shown in Fig. 1.
H Head ratio: HR = H/Ho

Head reduction factor: RH = 1- HR
Efficiency ratio: ER = / o
Ho Efficiency reduction factor: R = 1-ER
H
Water
Slurry
Fig. 1. Sketch defining the reduction in head and efficiency of a centrifugal pump pumping a solid
water mixture. H and H0 are heads in slurry and water service, respectively. The corresponding
efficiencies are η and η0.
A generalized diagram for estimation of the performance derating in head for settling
slurries is given in Figure 2, see ANSI/HI (2005) or Addie et al.(2007)
RH (%)
HR
Fig. 2 Generalised solids-effect diagram for pumps of various sizes (impeller diameters).For solids
concentration by volume 15% with relative density of solids, 2.65, and a negligible amount of fine
particles smaller than 75 micron. From ANSI/HI (2005).
Abritrary HR or RH values from Fig. 2 are then obtained through concurrent

corrections for solids concentration, relative density and content of particles smaller than
75 micron, for details see ANSI/HI (2005) or Addie et al. (2007).
Typical pump performance effects when pumping viscous Newtonian liquids are
shown Fig. 3.
Water viscosity
Higher viscosity
H
Flow rate
Fig. 3. Schematic sketch on pump characteristics depending on fluid viscosity
ANSI/HI (2004) provides a generalized derating method of the water performance

when pumping viscous Newtonian liquids (Fig. 3), i.e. when the property can be defined
by a constant viscosity value. It can be seen from Fig. 3 how both the efficiency and the
best efficiency region is influenced, indicating that the pump behaves as a smaller unit.
Correction factors for head, flow rate and efficiency are based on viscosity, head, flow
rate and rotary speed.
Non-settling type of slurries means that the liquid contains solids in the form of very
small particles, practically giving no segregation or settling tendencies during the
handling in pumps and pipelines. These mixtures behaves often in a non-Newtonian way
and exhibit a yield stress which is defined as the minimum stress required causing the
slurry to flow. Highly concentrated industrial slurries can often be approximately
represented by a yield stress and a straight line for larger shear strains in a rheogram. The
slope of the line is often denoted ‘tangent viscosity’. If the linear relationship prevails to
zero shear strain then the slurry is a Bingham type of medium, for which the slope
(tangent viscosity) is termed plastic viscosity or coefficient of rigidity.
Pump performance derating results for two non-Newtonian slurries fluids causing
unstable head curves are shown in Figs. 4 and 5.
Fig. 4. The effect of slurry on the pump head and efficiency at solids concentrations by volume of
31 and 35 % of a scrubber sludge product. Impeller diameter 0.63m. τy =35 Pa and η =0.03 Pas
for 35%. From Sellgren et al.(1999).
30
Head (m)
25
20
15
0 25 50 75 100
Flow rate (l/s)
Fig. 5.Kaolin, τy =240 Pa, η= 0.034 Pas. A 0.15 by 0.1m pump with a 5-vane metal impeller with
diameter 0.365m. BEP at about 50 l/s. From Kabamba (2007).
The flow rates for which the head cannot be maintained for the fine-particle slurries
in Figs. 4 and 5 are about 50% of the flow at the best efficiency flow rate (BEP).The
results in the figures show how the head diverges for low flow rates, which is opposite to
what is shown in Figure 3 for viscous Newtonian liquids. The deratings in Figure 3 are
caused by increased flow resistance within the pump due to an increased liquid viscosity.
The slurries in Figs 4 and 5 are composed of small solid particles which takes up 20 to
35% of the volume. Most of the kaolin particles are smaller than about 2 microns and the
suspension can be considered to be fully homogeneous.
The ANSI/HI (2005) slurry pump standard relates for non-settling slurries that
corrections can be based on a characteristic viscosity, for example the tangent viscosity
defined above or other viscosity concepts. With industrial non-settling slurries typical
deratings in head and efficiency may not exceed 5 to 10% when operating close to the
best efficiency point of the pump. However, with respect to the complex behaviour
sometimes observed for lower flows, it is suggested in the standard that the pump
manufacturer should be consulted for guidance regarding non-Newtonian effects on
pump performance.
1.1 MODELLING
Apparent viscosity is the viscosity at a particular rate of shear expressed in terms of a

Newtonian fluid. With a non-Newtonian fluid with a yield stress, the apparent viscosity
increases dramatically for small shear rates. Most modelling approaches of the
progressively derating with decreasing flow rates have been based on assumptions about
a representative shear rate within the pump. The corresponding apparent viscosity is then
used with the Hydraulic Institute Newtonian liquid method to calculate the derating
factors. Sellgren and Addie (1994) simply used the bulk shear rate 8V/D (V is velocity)
calculated for the pump suction pipeline diameter, D, to simulate the reduction in head in
Fig.4. The agreement was good for the conditions covered in Fig 4, however the
approach does not represent conditions in the pump entrance region and it could not be
validated when compared to other data, for example Walker and Goulas (1984).
Pullum et al. (2007) related an apparent viscosity to a characteristic geometric section

within the pump, thus representing a characteristic shear rate. The detailed section
geometry is so far mainly determined by experimental data, however results correspond
to realistic main pump dimensions and the approach seems promising for generalisations.
A different modelling approach by Wilson and Sellgren (2006),directly linked to the

yield behaviour simulated well the experimentally observed dramatic drop in pumping
head at flow rates less than about 50 % of BEP in Fig.4. They expressed a dimensionless
yield stress parameter to the actual flow rate over the flow rate at BEP. Build-up
tendencies of material within the pump have been documented in pictures with an acrylic
pump, Pullum (2006). These attached lumps can occlude fractions of the passages within
the pump which may seriously influence the head produced.
2. OBJECTIVES
The objective in this paper is to characterize the effect of solids on the performance of
centrifugal slurry pumps for various industrial slurries with a complex behaviour as
defined above. The effect on the performance is often moderate to low when the head
curves are fully maintained. Experimental results will be presented and various
mechanisms causing the effect of solids will be discussed.
3. CHARACTERISATION
Values of HR and ER in the range of 0.9 to 1 were presented by Cooke( 2007) from
results obtained for various high-density tailings products with yield stresses of up to
100 Pa and η from 0.01 to 0.07. He related the derating parameter to a chart that goes
back to Walker and Goulas (1984) and a pump Reynolds number.
Z Di2 U m
Re 1
K
where ω is rotational speed, Di, impeller size ,ρm slurry density and η and plastic or
tangent viscosity Cooke (2007) recommended that installations are not designed for Re<
3.105 without conducting test work. This value corresponds in the chart to about 0.8 and
about 0.9 for Re about 106. In this range the derating on head will be small with
HR>0.95.With Re in excess of 107 any derating can be neglected. One of the tailings
product investigated by Cooke had 70% of the particles larger than 75 micron with a d50
here estimated to about 0.1mm. Even if the settling slurry diagram in Fig.2 has not been
intended for this type of mixtures, it gave HR values of about 0.95, i.e. in accordance
with observations and Eq.1.
3.1 SIMULATED TAILINGS DATA
A simulated a tailings product was characterized by a fully sheared yield stress of

about 100 Pa., Addie and Sellgren (2007). The rheology was evaluated from the pipeline
loop results and expressed in terms of the yield stress and a tangent viscosity. The
corresponding tangent viscosity was estimated to 0.055 Pas. Resulting pump deratings
are given in Fig. 7.The pump was a modified 3-vane metal unit with an open shroud 0.3
m diameter impeller having a simple auger-like inducer.
16 80
14 70
12 60
Efficiency (%)
10 50
Head (m)
8 40
6 30
4 20
2 10
0 0
30 40 50 60 70 80
3
Flowrate (m /hr)
Head, Water Head, Slurry Efficiency, Water Efficiency, Slurry
Fig.7 Pump performance derating for a simulated tailings product characterized by a yield stress of
about 100 Pa. From Addie and Sellgren (2007).
It follows from Fig. 7 that the estimated head curve was stable HR and ER were
found to be 0.95 and 0.86, respectively. These values agreed reasonably well with the
Re-criterion in Eq.1 and the HI-method based on Newtonian fluids. Whitlock and
Sellgren (2002) reported tests with a paste-like slurry with solids concentration by
volume, C=38 % in the 0.3 m impeller pump described in connection with Figure 7,see
Table 1. It follows that values of HR and ER were 0.9 and 0.85.
Table 1
Resulting HR and ER for τy =240 Pa and η=0.07 Pas. Whitlock and Sellgren (2002)
Pump type,
Rotary Impeller
Product C HR ER comments and
speed diameter
reference
0.1 by 0.075 m open
Clay-
38 1300-2000 0.3 0.9 0.85 shrouded 3-vane
sand
impeller
3.2 KAOLIN AND CMC DATA
Bootle (2006) investigated the derating effects for high and low yield stress kaolin
slurries with C= 28 and 37% with a 4-vane semi open 0.244-m impeller pump. A
chemical additive was used to reduce the yield stress from 200 Pa to 7.5 Pa, see results in
Table 2.
Table 2
For C=28%, τy =200 Pa , η=0.14 Pas. For 37%, τy =7.5 Pa , η=0.018 Pas after adding a chemical.
Rotary Impeller Pump type, comments
Product C HR ER
speed diameter and reference
0.186 by 0.1 m semi-
28 0.99 0.83
Kaolin 1500 0.244 open 4-vane impeller,
37 0.99 0.82
Bootle (2006)
It follows from Table 2 that the performance is not affected despite the big change in
yield stress in this case and the values for HR 0.99 and ER 0.82/0.83. Any derating effect
was not observed on the head.
Results by Kabamba with a CMC fluid is shown in Fig.8. CMC is a non-Newtonian
liquid with pseudoplastic properties and no yield behaviour.
50
Head (m)
40
30
20
0 20 40 60 80 100 120
Flow rate (l/s)
Fig. 8. Head curve for a CMC fluid with pseudoplastic coefficients 3.69 and 0.598. Bingham
parametrar for large shear rates gave τy= 50 Pa and η =0.2Pas. Pump and pumping conditions as in
Fig.5. From Kabamba (2007).
The tendency of a lowering head in Fig.8 indicates that a yield stress not necessarily
is a determining factor for the development of an unstable head curve.
4. DISCUSSION
With the exception of Bootle’s high-yield stress results in Table 2, it was found that
head could not be fully maintained at lower flow rates as shown here for results from
various non-settling fine particle slurries in Figs. 4 and 5 and a yield-free power law
liquid in Fig. 8. These media show various degrees of non-Newtonian behaviour
expressed by an apparent viscosity that increases for decreasing flow rates.
The observed set-in of destabilisation effects was related to how far away from the
BEP-region the pump operates. The further to the left of BEP the more recirculation and
shock losses which basically increase with the square of the flow rate at BEP minus the
actual flow rate. An increase in rotary speed in this domain for a constant flow rate may
contribute to the destabilisation because the distance away from BEP increases, see
schematic example in Fig.9.
1.5 N B BEP- Line

Relative head
N
1
A
1 2
Relative flow rate
Fig. 9. An increase in rotary speed from A to B through an increase in rotary speed, N, for a
constant flow rate means operation further away from the BEP-region.
The head curve was stable for the data in Table 1 at C=38%. However, with C = 41 %
and a τy of about 350 Pa the situation in Fig.9 occurred when a tendency of a drop in
head was met with an increase in rotary speed. Detailed evaluation of experimental data
showed that HR dropped from 0.84 to 0.70 in the BEP-region when N increased from
1775 to 1925 rpm for practically constant values of flow rate, C, and pipeline friction
losses during about 7 minutes, Fig. 10.
1
HR
ER
0.9
HR and ER
0.8
0.7
0.6
1750 1800 1850 1900 1950
N (rpm)
Fig.10. Decrease in head and efficiency ratio for C= 41% ( τy about 350 Pa) when N increases with
practically constant flow rate and pipeline friction losses during about 7 minutes.
This situation may be caused by progressive build-up of material within the pump
with geometrical modifications of flow passages within the pump that promote losses and
affects the suction performance. With a constant flow rate the effect of the apparent
viscosity is not caused by a decreased flow rate, only of the change in local flow pattern
associated with the increase in rotary speed. Following Eq.1, an increase in N should
increase Re and thus contribute to a stabilisation. However, the situation is complex, the
cause of the increase in N may also be an increase in η. In addition, Sery et al. (2006)
reported from kaolin tests a slightly smaller drop in head when the rotary speed
increased.
The 4-vane semi open 0.244-m impeller pump used by Bootle (2006) for kaolin was
modified with an enlarged suction inlet from 0.1 to 0.186 m and integrated flow inducer
scoops, which project into the enlarged inlet. The suction performance is strongly
influenced by the detailed configuration of the entry of the impeller inlet with the vanes.
The rapid changes in direction along the flow path cause shock losses which together
with recirculation dominates the losses for low flow rates compared to the BEP-region.
Roudnev (2004) found that NPSHR can be several times larger than the water value
based on experimental results from several Russian references from 1972 to 1986 with
Bingham-type of slurries, for example heavy media magnetite products. The greater
requirement was related to blockage tendencies in the entrance region by the yield
behaviour resulting in higher velocities and thus lower pressure. It is not known if the
findings reported by Roudnev was associated with any unstable head curve for low flow
rates. However, the discussed mechanisms have some similarities to possible blockage
from vapour in the entrance region sometimes observed for liquid pumping at low flow
rates with strong local recirculation causing cavitational conditions, Girdhar and Moniz
(2004).
An unstable head curve at low slurry flow rates may be the combined effects of
recirculation with strong local eddy currents and vortexing together with a complex
solid-liquid mixture or fluid that magnifies the entrance shock losses to an extend that the
pump head cannot be maintained.
5. CONCLUSION
Various experimental results are discussed and found to give maximum deratings in
head and efficiency of about 10 and 15 %, respectively, provided stable head curves can
be maintained. Magnified shock losses, circulatory flows and blockage of slurry and
possibly vapour tendencies in the pump entrance region may be considered as
mechanisms behind unstable head curves together with the flow behaviour of the slurry.
REFERENCES
1. Addie G.R. and Sellgren A.2007.A first slurry pump standard and some
implications for paste applications. Proceedings of the Tenth International
Seminar on Past and Thickened Tailings 13-15 March 2007, Perth, Australia.
2. Addie G.R., Roudnev A.S. and Sellgren A.2007.The new ANSI/HI centrifugal
slurry pump standard, Proceedings, The 17 th Int. Conference on the Hydraulic
Transport of Solids, Cape Town, South Africa,pp.205-218
3. ANSI/HI 12.1-12.6. 2005.standard for rotodynamic (centrifugal) slurry pumps.
4. ANSI/HI 9.6.7. 2004. Effect of liquid viscosity on rotodynamic (Centrifugal and
vertical) pump performance 9.6.7,2004.
5. Bootle M.J., 2006. Practical aspects of Transporting Past with Rotodynamic
Slurry Pumps. Proceedings of the Ninth International Seminar on Past and
Thickened Tailings 3-7 April 2006, Limerick Ireland. pp. 413-427.
6. Cooke R. Thickened and paste tailings pipeline systems-Design procedure, part
2. Proceedings of the Tenth International Seminar on Past and Thickened
Tailings 13-15 March 2007, Perth, Australia.pp.129-140.
7. Girdhar P. and Moniz O. 2004: Practical centrifugal pumps. Design, operation
and maintenance. Elsevier.
8. Kabamba B.M. 2007.Evaluation of derating procedures for centrifugal pump
performance, Magister Technologiae thesis, Cape Peninsula University of
Tecnology, South Africa.
9. Pullum, L., 2006. Personal communication: courtesy of the Thermal and Fluids
Engineering Group, CSIRO, Australia, 2006.
10. Pullum, L., Graham.L, J, W., Rudman, M., 2007. A De-Rating Method for
Centrifugal Pumps Pumping “Paste”: Proceedings of the Tenth International
Seminar on Past and Thickened Tailings 13-15 March 2007, Perth, Australia.
pp. 175-184.
11. Roudnev,A.S.2004 Slurry pump suction performance considerations,Proc.16 th
Int. Conference on Hydrotransport, BHR Group,Santiago,Chile,pp137-150
12. Sellgren A. and Addie G.1994. Pump solids effect when handling non-settling
slurries, Proceedings, Hydro-mechanisation 9. Miskolc, Aug. 29-31,Hungary.
13. Sellgren, A., Addie, G. & Yu, W.C., 1999. Effects of non-Newtonian mineral
suspensions on the performance of centrifugal pumps, Journal of Mineral
Processing and Extraction Metallurgy Review, 1999. Vol. 20, pp 239-249.
14. Sery G., Kabamba B., Slatter P., 2006. Paste pumping with centrifugal pumps:
Evaluation of the Hydraulic Institute Chart De-Rating Procedures. Proceedings
of the Ninth International Seminar on Past and Thickened Tailings 3-7 April
2006, Limerick Ireland. Australian Centre for Geomechanics, Perth, 2006. pp.
403-412.
15. Walker, C.I. & Goulas, A.,1984. Performance characteristics of centrifugal
pumps when handling non-Newtonian homogeneous slurries, Proc. Instn. Mech.
Engrs. Vol. 198A, No. 1, 1984.pp. 41-49.
16. Whitlock L. and Sellgren A.,2002. Pumping a simulated paste-like sludge with a
modified centrifugal pump, Proceedings, The 7 th European Biosolids Organic
Residuals Conference, November 17-17,Wakefield, UK,2002.
17. Wilson K.C., and Sellgren A.,2006. Effect of strongly non-Newtonian slurries
on the head and efficiency of centrifugal pumps, Proceedings, 11th International
conference on transport & sedimentation of solid particles., Georgia, September,
2006.
18. Wennberg, T., Sellgren A. 2007. Pumping evaluations with paste tailings
thickened close to the surface disposal area. Proceedings,The 17th Int.
Conference on the Hydraulic Transport of Solids, Cape Town, South Africa.
Paper B
Pumping evaluations with paste tailings thickened close to the surface disposal area
Wennberg T and Sellgren A
The 17th international conference on the hydraulic transportation of solids in pipes,

Hydrotransport 17, 2007, Cape Town, South Africa
Thord Wennberg, Anders Sellgren
The South African Institute of Mining and Metallurgy
The 17th International Conference on the Hydraulic Transport of Solids ‘Hydrotransport 17’
PUMPING EVALUATIONS WITH PASTE TAILINGS
THICKENED CLOSE TO THE SURFACE DISPOSAL AREA
Thord Wennberg
Luossavaara-Kiirunavaara AB, Sweden
Anders Sellgren
Luleå University of Technology, Sweden
Abstract
Elevated location of a paste-thickener on a ridge close to the disposal area is considered at a Swedish iron
ore mine. About 0.7 Mtonnes of thickened tailings are planned to be layered as paste in the vicinity of the
thickener during several years with distribution up to 900m after about 20 years. In order to clarify the
pipeline pumping characteristics of the tailings product for volumetric solids concentration from 40 to
50%, experiments in loop systems with pipeline inner diameters from 0.063 to 0.152m have been carried
out with various types of pumps. The results and comparisons carried out here showed that the rheology
cannot simply be related to the performance of centrifugal pumps, factors related to pump design and
built up tendencies of material within the pump and other particle properties may be of equal importance.
During the first years of operation in the considered application only one thickener underflow pump is
required. Therefore the reliability and performance for various pump configurations are planned to be
systematically investigated in order to meet the most effective long-term solution for paste thickener
operation, pipeline distribution and placement in the disposal area.
Introduction
Luossavaara-Kiirunavaara AB (LKAB), an iron ore company with mines in the northern part of Sweden
is continuously considering new technologies for the handling, transportation and disposal of waste rock
and tailings. In the LKAB concentrator in Svappavaara about 0.7 Mtonnes of dry tailings per year is
transported nearly 1 km as a slurry in parallel pipelines connected to a flume discharging into a tailings
pond. In order to limit the cost of tailings management in the existing area surface disposal of a highly
thickened slurry at slopes of 5% has been considered. The thickening may take place adjacent to the plant
or close to the disposal area about 1.5 km away, Fig.1, Wennberg and Sellgren (2005).
The location of the thickener is mainly a balance of the cost for high pressure pumping and the
disadvantage of having the thickening facility remote from the rest of the processing taking place in the
concentrator. Today, it seems that a location of the thickening at 2 in Fig.1 is the most attractive
alternative.
Various design features related to the underflow pump, the downstream distribution system and the
geotechnical behaviour when the paste is in place are in the literature coupled to characteristic paste yield
stresses. Values related to “unsheared” and “fully sheared” conditions goes back to the effect of applied
shear and exposure time, i.e. the shear history. As will be discussed later there are some uncertainty about
shear history effects.
Objective and Scope
The objective here is to present experimental results from various loops with inner diameters of 63, 75,
101, and 152mm with centrifugal and positive displacement pumps, the latter of screw or piston type.
These diameters are expected to be representative for various parts of the distribution system. Due to the
thickener location close to the disposal area, the discussion will be focused on the performance of
centrifugal pumps when handling high concentration slurries.
Pumping evaluations with paste tailings thickened close to the surface disposal area 1
C
+ 369 1
+ 436
+ 411
2
Emergency
tailings line
+ 342
W + 359
S
+ 365 N
Legend D
C: Concentrator 0 0.5 1 km
D: Disposal area
S: Settling pond Thickened tailings
W: Clarification pond Tailings
Water
Figure 1. Schematic sketch of the two alternative locations of the high-density thickener.
It was not possible to do on-site large-scale pipeline pumping testing and direct shear history observations
at the time of pilot-scale paste generation because of low production rates. Therefore, the pipeline loop
tests were carried out with the actual thickened product at suitable locations with stored, transported and
re-mixed pastes. During the first years of operation in the considered application only one thickener
underflow pump together with a short distribution pipeline is required. Therefore questions raised in this
study and shear history effects are planned to be systematically investigated on-site in order to meet the
most effective long-term solution for paste thickener operation, pipeline distribution and placement in the
disposal area.
Experimental Study
The average particle size distribution was 25 % less than 10 microns with average particle size, d50 ,about
27 microns and with about 80% of particles finer than 90 microns. Figure 2 shows schematically the test
loop procedure, most often with segments with two pipeline diameters where simple gauges were
installed for pressure drop readings. The gauges were at times mounted close to the pump and sometimes
a soft pipeline bend was included in the pressure readings. The flow rate was simply measured with a
barrel and a stop watch and the slurry density was obtained from a volumetrically calibrated tank and
dried samples.
Barrel for
flow control
4m 50 m 4m
Figure 2. Schematic description of one of the used loop systems. Truck feeding of a displacement
pump in a 120 m long loop with 0.15- and 0.10m- diameter pipelines.
The measured pressure gradient dp/dx along a pipe is related to the wall shear stress w and the friction
loss gradient j (m slurry/m) through the following relationship.
(dp / dx) D gj
gjD
w (1)
4 4
where and D is the slurry density and pipeline diameter, respectively.
The performance of a centrifugal pump is affected when handling slurries. The relative reduction of the
clear water head and efficiency for a constant flow rate and rotary speed may be defined by the ratios and
factors shown in Figure 3.
Head ratio: HR = H/H0

Head reduction factor: RH = 1- HR
Efficiency ratio: ER = 0
Efficiency reduction factor: R = 1-ER
Figure 3. Sketch defining the reduction in head and efficiency of a centrifugal pump pumping a
solid water mixture. H and H0 are heads in slurry and water service, respectively. The
corresponding efficiencies are η and η0.
In order to compare pump performance results with various rotary speeds, n, and literature data,
dimensionless head and flow rate parameters were used, see for example Wilson et al. (2006).
gH Q
Head: 2
Flow rate: 3
(2)
n 2 Di nDi
where H is the head and Q is the flow rate and Di is the impeller diameter.
Results
Pipe wall stress results (Eq.1) versus the viscous scaling parameter 8V/D (V is velocity) for various
diameters and volumetric solids concentration ,C= 40.4%, are shown in Fig. 4 and compared to larger
diameter data from Saurmann (1982) for a tailings product with similar particle size distribution as used
here.
Results for various diameters up to 0.152 m for C-values of about 48 and 50% are shown in Fig. 5 where
direct yield measurements also were available for 48 %. This is the value considered for pipeline pumping
design which here has been estimated to be on the conservative side with respect to paste quality
requirements.
The line in Fig. 5 for 48% is considered to represent an average representative design shear stress. For
example, with D=0.15 m and 92 tonnes/h (0.7 Mtonner/y) then the flow rate is 68 m3/h with V about 1.1
m/s and 8V/D = 57 s-1 giving w about 200 Pa in Fig.5. The corresponding pressure requirement from
Eq.1 will be 530 kPa/100m or head losses of about 30m slurry per 100m horizontal pipeline.
600
Pipeline diameter
0.063 m
500
(Pa)
0.076 m
0.101 m
400
w
0.2 m, Saurmann (1982)
Shear stress
300
200
100
0
0 100 200 300
8V/D (sec-1)
Figur 4. Shear stresses versus the rheological scaling parameter 8V/D from pipeline test whith a
concentration of 40.4 per cent by volume.
600
Pipeline diameter
50 % 0.063 m
500
(Pa)
0.076
0.101 m
400
w
0.152 m
Vane-type viscometer
Shear stress
300
47.7 %
200
100
0
0 100 200 300
8V/D (sec-1)
Figur 5. Shear stresses versus the rheological scaling parameter 8V/D from pipeline test with
solids concentrations by volume of about 48 and 50 %.
The pipeline results in Figs. 4 and 5 were based on rough measurements and represent re-mixed and
mildly to strongly circulated slurry or a product that had been stored for one to several weeks and then
brought into a loop system. It was not possible to evaluate any obvious rheological trends related pipeline
diameter, age or exposure time in the loop system.
Operation in a 40 m long 0.065 m diameter hose was also used for indicative performance measurements
and observations for a centrifugal pump with a 0.065 m-suction and a 0.05 m-discharge diameter with a
4-vane semi-open impeller with diameter 0.23 m. A few pump efficiency values ( ) were estimated from
the variable speed unit. The flow rate at the start corresponded to a velocity of 1.55 m/s at 1480 rpm for
C=45.5% with the intention to successively increase the concentration at approximately the same flow
rate.
With C about 48%, it was not possible to maintain the flow through increased rotary speeds, however the
system operated for a long time in a semi-stable way at extremely low velocities of 0.1 to 0.3 m/s for
rotary speeds of 1840 to 1940 rpm.
0.18 100
0.16 H
H Slurry
Slurry
Head water
0.14 Slurry
h Slurry
0.12
2
gH/n Di
(%)
0.1
2
Efficiency water
50
0.08
0.06
0.04
0.02
0 0
0 0.002 0.004 0.006 0.008 0.01
3
Q/nDi
Figur 6. Dimensionless representation of high concentration slurries derating result.
Discussion
Indicative measurements at an early stage with gravity flow in a 11 m long 25mm-diameter hose from a
pilot scale thickener at C=44% showed a yield stress slightly above 100 Pa, which make sense in
comparison to the results in Figures 4 an 5 for about 40 and 48%,respectively, and the scatter related to
the various testing conditions.
Parts of the underflow in paste thickeners are today often circulated back into the base through pumping
in order to agitate the paste. Pilot scale thickening and trench deposition tests for a similar tailings product
as investigated here with and without circulation with a centrifugal pump showed that the agitated paste
flowed more easily down the slope resulting in a lower deposition angle, Engman et al. (2004)
The thickened and flocculated slurry flows by gravity in the base of the thickener to the pump.
Centrifugal pumps are known to impart high shear intensities resulting in a reduced slurry yield stress.
Furthermore, the slurry is exposed to aging and various degrees of shear in the downstream pipeline
distribution system. Published results are difficult to evaluate and sometimes contradictory. For example,
Schaan et al. (2005) reported for a flocculent-dosed coarse sand-silt/clay tailings paste that the dominating
relative yield stress reduction took place in the 300 m long 0.075 m-diameter pipeline after two
centrifugal pumps where the yield stress decreased about 50%. An 80% decrease in yield stress through
pump passages for a red mud tailings slurry with no pipeline effect was reported by Sofra (2006).
Arbuthnot and Triglavcanin (2005) found that a centrifugal pump in the underflow approximately halved
the yield stress with a gold tailings paste. They found a similar effect for an underflow circulating
arrangement in the thickener. High pressure systems with positive displacement pumps are normally also
fed with centrifugal pumps. Therefore, independent of the system configuration, slope forming
mechanisms can be considered to be related to shear stresses that are much lower than those observed in
the thickener underflow.
The reported shear history experiences and the observation of a slope dimishing effect related to paste
circulation in the thickener, raise questions related to system effectiveness and flexibility. With a long-
term fixed overall slope requirement in the disposal area, flexibility can be promoted with a split of the
pipeline flow into several streams just before the discharge end, giving higher slopes following an
indicative criterion, Engman et al. (2004). In order to plan for investigational procedures related to system
effectiveness during the first years of operation with simple underflow pumping arrangements, the initial
choice of equipment will be important. Therefore the rest of the discussion will be focused on centrifugal
pump performance derating when handling high concentration slurries.
The pump performance results in Fig.6 addresses the problem of maintaining steady state operation when
pumping a paste with various solids concentrations. Walker and Goulas (1984) presented results with an
unstable head curve for coal and kaolin clay slurries with yield stresses of about 8 Pa. Results for a four-
vane closed impeller with diameter 0.352 m and a constant rotary speed of 1200 rpm are shown in Figure
7.
0.18 100
0.16 Head water
0.14
0.12 Efficiency water

2
gH/n Di
0.1
2
50
0.08
0.06
0.04
0.02 Coal, C =31%

Kaolin, C =29%
0 0
0 0.001 0.002 0.003 0.004
3
Q/nDi
Figure 7. Dimensionless representation of coal and kaolin derating results from Walker and Goulas
(1984). Filled and unfilled markings correspond to head and efficiency, respectively.
The derating effect in efficiency was larger for the coal which had a similar size distribution as the
tailings product here, while the kaolin clay had an average diameter of about 1μm. In addition to head and
efficiency reductions in the best efficiency-region (bep), the head could not be maintained for lower flow
rates, thus giving an unstable head curve. The tendency with a drop in head for lower flow rates for
constant rotary speeds and solids concentrations for non-Newtonian slurries has also been observed with
larger pumps, see Fig 8.
Further kaolin results at C about 30% by Walker and Goulas (1984) with another smaller pump with 5-
vane closed or 3-vane semi-open impellers with diameters of 0.24 m showed small or negligible effects
on the head and ER about 0.8. Basically the same pump with a 4-vane semi-open 0.244-m impeller,
modified with an enlarged suction inlet from 0.1 to 0.186 m with integrated flow inducer scoops which
project axially into the enlarged inlet, has recently been tested for high and low yield stress kaolin
slurries, Bootle (2006). The results discussed above have been brought together in Table 1 and Figure 9
where also derating data from Fig.8, and for a simulated paste-like slurry in a 0.3 m-impeller pump with
an auger-like inducer operating with a stable head curve, Whitlock and Sellgren (2002).
Figure 8. The effect of slurry on the pump head and efficiency at solids concentrations by volume of
31 and 35 % of a scrubber sludge product. Impeller diameter 0.63m. From Sellgren et al.(1999).
Table 1.Some derating results available in the literature. Solids density about 2650 kg/m3 and 1450
kg/m3 for the coal. Number 4 comprised an auger-like like inducer and 2 an inducer in an enlarged
inlet. C in % and rotary speed in rpm with impeller diameter in m.
Rotary Impeller Pump type, comments and
Notation Product C HR ER
speed diameter reference
0.1 by 0.075 m closed 4-
1.1 Coal 31 0.91 0.63
1200 0.352 vane impeller, Walker and
1.2 Kaolin 29 0.92 0.87
Goulas (1984)
0.186 by 0.1 m semi-open 4-
2.1 28 0.99 0.83
Kaolin 1500 0.244 vane impeller, Bootle
2.2 37 0.99 0.82
(2006)
0.2 by 0.15 closed 4-vane
3.1 35 0.91 0.85
Sludge 1200 0.63 impeller, Sellgren et al.
3.2 31 0.97 0.93
(1999)
0.1 by 0.075 m open
Clay-
4 38 1300-2000 0.3 0.9 0.85 shrouded 3-vane impeller,
sand
Whitlock & Sellgren (2002)
400
350
2.1
(Pa)
300
4
w
250
Shear stress
Fig. 5, C=47.7
200
150
100
3.2 1.1
50 3.1
1.2
0
2.2
0 100 200 300 400 500 600 700 800 900 1000
-1
Shear rate (s )
Figure 9.Approximative rheological results for the slurries listed in Table 1 and for C=48% (Fig.5).
The result 2.1 and 2.2 in Table 1 and Figure 9 were characterized by yield stresses of 200 and 1.8 Pa,
respectively, corresponding to ER-values of 0.82 and 0.83. Any derating effect was not observed on the
head. The efficiency drop was from 60% for water to about 50% for slurry. Thus the deratings were
practically identical and independent on the large difference in rheological behaviour. The trend in kaolin
results at C=29%, with a yield stress of about 8 Pa(1.2, Fig.9), from Walker and Goulas 16 years earlier in
the corresponding standard pump was the same as observed for 2.2 in Figure 9, with a nearly negligible
yield stress. Tendencies of turbulent flow can be seen for 3.2 in Fig. 9.
For # 4 in Table 1 and Fig.9 at C = 38% the head curve was stable. However, with C = 41%, evaluation of
detailed experimental data provided by Whitlock (2006), showed that HR dropped from 0.84 to 0.70 in
the bep-region when N increased from 1775 to 1925 rpm for practically constant values of Q, C, and
pipeline friction losses during about 7 minutes, Fig.10.
1
HR
ER
0.9
HR and ER
0.8
0.7
0.6
1750 1800 1850 1900 1950
N (rpm)
Figure 10. Decrease in head and efficiency ratio for C= 41% when N increases with practically
constant flow rate and pipeline friction losses during about 7 minutes.
This situation may be caused by progressive built up of material within the pump with geometrical
modifications of flow passages within the pump that promote losses. Cavitation was found to be a less
likely explanation even if the detailed suction performance of the modified pump was not known. Built
up tendencies of material within the pump have been documented in pictures with an acrylic pump,
Pullum (2006). These “attached” lumps can occlude fractions of the passages within the pump which may
seriously influence the head produced, for example at low Q as seen in Fig.8. The experimentally-
observed dramatic drop in pumping head at flow rates less than about 50% of bep for # 3 in Table 1 was
linked to the yield behaviour in a modelling approach by Wilson and Sellgren (2006).
The results and comparisons carried out here showed that the rheology cannot simply be related to the
performance of centrifugal pumps, factors related to pump design and built up tendencies of the paste
material within the pump and other particle properties may also be of importance.
The results with the modified pumps for 2.1 and 4 in Table 1 and Fig. 9 with a yield stress exceeding
about 200 Pa showed stable steady-state operation. Based on this a modified pump may be required for
the tailings considered here. Furthermore, ER-values of 0.80 to 0.85 can be expected. However, the
materials discussed in Table 1 and Fig.9 were not pastes or tailings products. The coal with ER=0.65 had
a similar size distribution as the tailings considered here.
Laboratory pump performance results with a 0.15 by 0.1 m 4-vane rubber lined pump (impeller diameter
0.45 m) operating with a stable head curve at C=36% are available by Sellgren and Väppling (1986). The
size distribution was similar to the product studied here and the rheology at C=36% was close to 3.1 in
Fig.9. Provided the head curve remains stable, then HR and ER were here extrapolated for the actual C=
48% to 0.8 and 0.65, respectively.
The results in Fig. 9 have one rheological parameter in common, the tangent viscosity, i.e. the slopes of
the lines, with values from 0.04 to 0.14 Pas. As indicated by Walker and Goulas (1984) and others, this
viscosity may be used together with the Hydraulic Institute’s performance derating method for viscous
Newtonian liquids, see ANSI/HI(2004),ANSI/HI(2005).These viscosity values would here give ER-
values from 0.7 to 0.9 for a suitable pump.
Conclusions
The overall pipeline loop results for the tailings slurry showed that the slurry friction losses can be
approximately characterized by a pipe wall shear stress of about 200 Pa, for example resulting in a
pressure requirement of about 530 kPa/100m or head losses of about 30 m slurry per 100 m pipeline in a
0.15 m-diameter pipeline for a solids concentration by volume of about 48%.
The results and comparisons discussed out here indicated that a modified type of centrifugal pump may
be required to avoid the risk of an unstable head curve. Furthermore, with an ER-value of 0.80, the total
pump efficiency including motor and drive may be about 40%.
It was found that the rheology cannot simply be related to the performance of centrifugal pumps, factors
related to pump design and built up tendencies of paste material within the pump and other particle
properties may be of equal importance. During the first years of operation when only one thickener
underflow pump is required, the reliability and performance for various types of pumps can be
systematically investigated in order to meet the most effective long-term solution for distributing the
paste.
Acknowledgements
Boliden Mineral AB and Petra Brodin, process engineer, are acknowledged for the cooperative use of the
loop arrangements in Fig. 2 in connection with underground hydraulic fill test work at the Garpenberg
mine, Sweden.
REFERENCES
ANSI/HI 12.1-12.6.standard for rotodynamic (centrifugal) slurry pumps
ANSI/HI 9.6.7. Effect of liquid viscosity on rotodynamic (Centrifugal and vertical) pump performance
9.6.7,2004
ARBUTHNOT L M. AND TRIGLAVCANIN R. M. Designing for paste thickening testwork and sizing
for paste thickeners, Paste 2005 Int’l seminar on Paste and Thickened Tailings, Santiago (Chile), 2005,
20-22 April, 2005. pp99-116
BOOTLE M.J., Practical aspects of Transporting Past with Rotodynamic Slurry Pumps. Proceedings of
the Ninth International Seminar on Past and Thickened Tailings 3-7 April 2006, Limerick Ireland.
Australian Centre for Geomechanics, Perth, 2006. pp. 413-427.
ENGMAN, M., A.SELLGREN, Å. SUNDQVIST, I.GOLDKUHL AND T. WENNBERG : Users

perspective on the design of high density base metal tailings systems, Proceedings of the11th Int.
Conference on Tailings and Mine Waste’04, Vail, CO, U.S.A, October, 2004,A.A. Balkema Rotterdam.
PULLUM, L., Personal communication: courtesy of the Thermal and Fluids Engineering Group, CSIRO,
Australia, 2006.
SAUERMANN, H.B., The influence of particle diameter on the pressure gradients of gold slimes
pumping, Proc. Hydrotransport 8, BHRA Fluid Engineering, Cranfield, UK, 1982. pp 241-248.
SCHAAN J, SANDERS, R S GILLIES R G,MCKIBBEN M J,LITZENBERGER C,SUN R-J,SHOOK C

A. Effects of shear history on the flow properties of flocculant-dosed thickening tailings slurries. HY15,
Banff, Canada, 2002
SELLGREN, A., ADDIE, G. & YU, W.C., Effects of non-Newtonian mineral suspensions on the
performance of centrifugal pumps, Journal of Mineral Processing and Extraction Metallurgy Review,
1999. Vol. 20, pp 239-249.
SELLGREN A.AND VÄPPLING L.,Effects of highly concentrated on the performance of centrifgal

pumps,proc.Int.Symposium on slutrry flows,FED,Vol 38,ASME,U.S.A. 1986.,pp143-148
SOFRA F. Rheological assessment-A road map for plant designers and operators Seminar on Past and
Thickened Tailings 3-7 April 2006, Limerick Ireland. Australian Centre for Geomechanics, Perth, 2006.
pp. 13-24.
WALKER, C.I. & GOULAS, A, Performance characteristics of centrifugal pumps when handling non-
Newtonian homogeneous slurries, Proc. Instn. Mech. Engrs. Vol. 198A, No. 1, 1984.pp. 41-49.
WILSON, K.C. & SELLGREN, A., Hydraulic transport of solids. Pump Handbook, 3rd Edition,
McGraw-Hill, 2001. pp. 9.321-9.349.
WHITLOCK L. AND SELLGREN A. Pumping a simulated paste-like sludge with a modified centrifugal
pump, Proceedings, The 7th European Biosolids Organic Residuals Conference, November 17-
17,Wakefield, UK,2002
WHITLOCK L. Personal communication, The Hydraulic Laboratory, GIW Industries Inc. 2006. U.S.A.
WILSON, K. C., ADDIE, G.R., SELLGREN, A. & CLIFT, R,. Slurry Transport Using Centrifugal
Pumps, 3nd Ed., Springer, New York, U.S.A, 2006.
WILSON K.C., AND SELLGREN A., Effect of strongly non-Newtonian slurries on the head and
efficiency of centrifugal pumps, Proceedings, 11th International conference on transport & sedimentation
of solid particles., Georgia, September, 2006.
WENNBERG, T., SELLGREN A. AND SUNDQVIST A., Feasibility considerations of high density
disposal at the LKAB Svappavaara concentrator, Proceedings, The international conference: Securing the
Future 2005,Skelleftea, Sweden, 2005. June 27-July 1
Paper C
Rheological and Depositional characterisation of paste-like tailings slurries
Wennberg T, Sellgren A and Goldkuhl I
Proceedings, Conference in Minerals Engineering, February 3-4 2008, Luleå, Sweden

RHEOLOGICAL AND DEPOSITIONAL CHARACTERISATION OF PASTE-LIKE
TAILINGS SLURRIES
Thord Wennberg1, Anders Sellgren2, Ingemar Goldkuhl3
1
Luossavaara-Kiirunavaara AB,
2
Luleå University of Technology,
3
MultiServ AB
Abstract
The rheological and depositional character of tailings from two base metal concentrators are
presented based on results obtained in connection with pilot-scale thickening tests and
feasibility considerations. Thickening to a volumetric solids concentration of over 45% ( 70%
by mass) close to the disposal area results in a non-segregating slurry that can be deposited at
average slopes of 3-4%.Operating data from reported installations are briefly described and
some macroscopic depositional mechanisms and design features are discussed.
1. Introduction
The feasibility of a high degree of tailings slurry thickening and deposition at angles of 3-4%
is determined by the possibility to fill more volume per unit surface area, thus limiting water
holding dam facilities and areas that need to be rehabilitated. The solids concentrations by
volume must often exceed 45% which means slurry properties that gives a conceptually even
slope with no segregation of particles and virtually no drainage of water. The term paste is
used in a slightly loose way here, it was originally associated with bauxite residue storage and
tailings backfilling in underground mines. The local conditions strongly determine the
technical-economical feasibility and environmental effectiveness of different degrees of
thickening and surface disposal.
With a paste system circulation of large quantities of water through the disposal area can be
avoided. Water availability may be a critical issue in arid or semi-arid areas where water
exposed in the disposal area is largely lost through evaporation. In sub-arctic and arctic
regions heat recovery with rapid water circulation through thickening can improve the
energy effectiveness in the mineral processing, see for example Hollow et al. (2007),
Oxenford and Lord (2006).
Disposal of tailings at high solids concentrations often means unsaturated conditions with
susceptibility to oxygen exposure in the comparatively small tailings storage areas. Aspects
related to deposition management schemas and the adoption of multi-layer capping for
tailings containing various degrees of AMD-generating minerals will not be covered here.
General aspects and experiences of using paste deposition for AMD-generating tailings can be
found in Jewell and Fourie (2006).
2.Objective and scope.

The objective is here to briefly characterize the rheological and depositional behaviour of two
tailings products based on test work and feasibility considerations at LKAB’s Svappavaara
concentrator and at MultiServ AB’s slag recovery plant. The aim is also to present some
depositional data from installations described in the literature and to discuss some
macroscopic depositional mechanisms.
The depositional characterization is mainly a summary of results presented in Engman et al.

(2004), where laboratory and field work are described in details together with results from
the thickening pilot-scale trials. The average particle sizes were from about 25 to 50 μm with
maximum sizes of about 350 μm. The rheological characterization is based on results from
pipeline flow in small-diameter hoses and flume flows ( Engman et al. 2004) and pipeline
loop pumping experiments in diameters from 0.060 to 0.2 m, Wennberg and Sellgren
(2007).Details on various considered equipments for thickening and pumping are not covered
in this study.
3.1Svappavaara
In the Luossavaara-Kiirunavaara AB ( LKAB) concentrator in Svappavaara close to the arctic
circle about 0.7 Mtonnes of dry tailings per year (90 tonnes/h)is transported nearly 1 km as a
slurry in parallel pipelines connected to a flume discharging into a tailings pond. In order to
limit the cost of tailings management in the existing area surface disposal of a highly
thickened slurry is considered. The need for costly dam rises in the disposal area are mainly
avoided with thickened disposal. The most attractive location of the thickener is on a hill
close to the disposal area (Wennberg and Sellgren, 2007),see Figure 1.
C
+ 369 1
+ 436
+ 411
2
Emergency
tailings line
+ 342
W + 359
S
+ 365 N
Legend D
D: Disposal area
S: Settling pond Thickened tailings
W: Clarification pond Tailings
Water
Figure 1. Schematic sketch of the two alternative locations of the high-density thickener at the
LKAB Svappavaara concentrator.
3.2 MultiServ AB
In the companys plant in central Sweden for recovery of metal slag about 0.23 Mtonnes per
year (50 tonnes/h) are planned to be thickened in the disposal area about 500 m away from the
concentrator. The thickened tailings will, depending on location of the thickener, then be
pumped 50 to 500 m to the disposal area, see Figure 2.
Figure 2. Schematic sketch of the disposal area at MultiServ AB in central Sweden.
During the first years of operation the existing disposal area will be covered with paste. This
means a covering of the disposal with simplified rehabilitation and also reduced leaching
from the deposit. In addition the ground water table will be lowered in the existing deposit
before the final tight seal capping.
The use of paste disposal has been chosen due to economical, environmental and safety
issues. The location in a valley downstream an existing deposit has been seen very favourable
for a paste system. The final location of the thickener has not yet been decided, but due to
high costs for paste pumping a location and the edge of the existing deposit is seen as the best
option. According to a newly received permit for the metal recovery operation including the
deposition and storage system, the paste system must be in operation before the end of 2009
and the new deposit latest 2014.
4 Characterization
Various design features related to the paste thickening, pumping and placement are in the
literature normally related to the rheology of the highly concentrated mineral particle-water
mixture, especially the yield stress. The yield stress is defined as the minimum stress required
causing the solid-liquid mixture to flow. The standard “slump” method for consistency
measurements in the concrete industry has been adopted and modified for paste-like tailings
slurries, see for example Jewell and Fourie (2006).
Early experiences of surface deposition and handling at very high solids concentrations
often comprised bauxite processing, i.e. clay minerals. Tailings from the diamond and oil sand
industries often also deal with active clay minerals. High clay content normally has a strong
effect on the rheological properties of the tailings slurry. For ground rock base metal mine
tailings, the finer (clay-sized) particles are not really clay minerals but rock flour, Fourie
(2002), a dissimilarity also marked by Robinsky (1999).Recent work with slurries of the type
considered here has shown that they may not behave in a completely non-settling way flowing
homogeneously during transportation, see for example Pullum (2007).
For simplicity, the characterization of the flow behaviour during transportation is here
adopting the generally accepted rheological parameters to describe the flow resistance of
homogeneously flows of paste-like tailings slurries.
4.1 Depositional characterization

The discharge from the pilot-scale thickening tests at both plants were with 0,025 m-plastic
hoses, at MultiServ AB with discharge in a trench with bottom width of 1,5 m and height
about 1 m. The deposition observations at the Svappavaara concentrator took place on a
horizontal floor with created cones with diameters of 7 to 8 m and heights of 0.5 to 1 m. In
addition, batch tests were carried out in a flume with width, height and length of 0.2 , 0.1 and
1.5, respectively. The various laboratory and field set-ups and detailed results are given by
Engman et al. (2004), a summary of deposition slopes are shown in Figure 3.
30
Floor
Batch
25 Trench
Trench, circulated
Batch, circulated
20
Slope (%)
15
10
0
38 45 50
C (%)
Figure 3. Observed deposition angles from 0.025m diameter discharge pipelines and batch
flume tests. Unfilled and filled marks correspond to the tests at LKAB and MultiServ AB,
respectively. Solids density 3000 kg/m3.
Circulated in Figure 3 is coupled to tests with agitated slurry through a circulating pump
arrangement in the lower portion of the thickener. It was observed that this slurry flowed
more easily down the slope. It follows from Figure 3 that average slopes of 10 to 15% were
obtained from a volumetric concentration of about 45% and that pump circulation in the
thickener at MultiServ AB tended to lower the slope, thus affecting the behaviour of the
slurry.
4.2 Rheological characterisation

Observations of continuous flow over a length of about 1m in a flume with width and height
0.2 and 0.1 m, respectively, were carried out at slopes from about 12 up to 27% where
approximately uniform conditions with a nearly constant depth were observed. No deposition
of particles were observed. The corresponding analysis for estimation of average shear
stresses is given in details by Engman et al. (2004).
The wall shear stress, τw, for the free surface flow in a flume or in a pipeline is related to the
pressure gradient, Δp/Δx, and the friction loss gradient, j , in m slurry per m through the
following relationship:
('p / 'x) D UgjD

Ww (1)
4 4
where ρ is the slurry density. D is here the hydraulic diameter which corresponds to the
pipeline diameter for flow in a filled pipeline section. For the free surface flume flow, D =
4A/P where A is the average section area and P is the average wetted perimeter.
Summarized pipe and flume wall shear stress results (Eq. 1 ) versus a viscous scaling
parameter, the bulk shear rate, 8V/D (V is velocity), are shown in Figure 4 covering pipeline
diameters of 0.025 and 0.05 m. Similar results are shown in Figure 5 for separate larger
pipeline diameter tests with Svappavaara-tailings. With large diameters, it was not possible to
do on-site large-scale pipeline pumping tests at the time of pilot-scale thickening tests because
of low paste production rates. Consequently, the tests were carried out with stored and re-
mixed pastes.
1000
Pipeline
Flume
Shear stress (Pa)
100
10
1
0.1 1.0 10.0 100.0 1000.0
8V/D (s-1)
Figure 4. Shear stresses versus the rheological scaling parameter 8V/D.From Engman et al
(2002).
1000
Shear stress (Pa) 100
Pipeline diameter
0.063 m
10
0.076
0.101 m
0.152 m
1
0.1 1.0 10.0 100.0 1000.0
8V/D (s-1)
Figure 5. Shear stress versus 8V/D for pipeline tests with various diameters at an average
solids concentration by volume of about 47%
The results in Figure 4 were based on indicative measurements (Engman et al. 2004) and the
accuracy of the data in Figure 5 were discussed in some details by (Wennberg and Sellgren
(2007).The uncertainties expressed in Figures 4 and 5 may to some extent express the
variability in practice, related local slurry properties and thickener operation. A perspective on
the results in Figures 4 and 5 may be represented by evaluated wall shear stresses of 500 and
20 Pa corresponding to concentrations of 50 and 40%,respectively.
5.Some case studies

Operating data focused on depositional conditions are briefly described based on information
given by Jewell and Fourie ( 2006) and Oxenford and Lord (2006).Solids concentrations by
mass are given within parenthesis together with the volumetric values.
The Kidd Creek mine and processing operations in central Ontario in Canada have a well
functioning thickening facility since 1995 where underflow at a concentration by volume, C,
of about 35% (62.5%) is pumped 1100 m in centrifugal pumps to a centrally located discharge
point. The average slope is 2.5 to 3 % and the average particle size, d50, about 20 μm and the
solids density 3100 kg/m3.
Surface disposal of paste was chosen to extend the life of the Myra Falls mine at the Pacific
coast of British Columbia in Canada. A combination of tank thickening with a portion
dewatered further in filters gave an average C-value of 36% (67%) The max. particle size
was about 100 μm and d50 about 12 μm with 20% smaller than 10 μm. Solids density=3600
kg/m3. Transportation with PD-pump with down-valley discharge resulted in a average
slope of about 3%.
In the Cabriza gold-copper mine in Peru about 200 t/h tailings are thickened to about 45%
(75%) in a deep cone thickener. The slurry becomes non-segregating at about 41% (70%).
Solids density 3700 kg/m3 and d50 from 15 to 25 μm. Deposition from a pipeline laid out
along a hillside with 10 spigots each about 0.1 m in diameter. Centrifugal pumps used so far,
positive displacement pump installed for extended pumping lengths.
The copper-gold Osborne mine operations in Australia produces about 200 tonnes/h of a
tailings product with 10% less than 20 μm,d50 about 70 μm and max. particles of 1 to 2 mm.
Average solids density= 3500 kg/m3. Deposition from an elevated central discharge point
means an advancing cone of deposited tailings. Full scale trials showed that average slopes of
1.1 and 2.5% were obtained for C-values of 26 (55%) and 35% (65%) respectively, when
deposited conventionally. A C-value of 42% (72%) was required for segregation-free
deposition and a slope of 4% was obtained for 44%(73%)
The Century zinc mine in Australia uses down-valley deposition for about 500 tonnes/h. The
intention here was not to go for the steepest slope for the tailings with d50= 12μm and solids
density 2820 kg/m3.A high rate thickener increased the C-value to about 30%(55%) giving
slopes of 1% and 0.6% in the upper and lower portions of the disposal area, respectively,
which was considered sufficient here.
6.Deposition mechanisms
6.1 Conventional deposition
In conventional low concentration tailings slurry discharge from a pipeline or flume in the
disposal area, the slurry flow takes place in self-forming meandering channels as the
particles deposit along the surface (“beach”).Segregation of the coarser and finer particles
occurs with coarser particles settling out close to the discharge point with finer particles being
transported further along the beach. The location of the channels change over time and the
detailed local directions is driven by gravity along the path of least resistance. The cross-
sectional shape of a channel that forms in underlying deposited tailings is a mechanism that
cannot be considered to be fully understood , ASCE (1998).The free surface flow often ends
at an open water surface (pond),where remaining fine particles settle out under water.
The continuously decreasing inclination of the deposited tailings beach profile normally
results in a slope that is less than 0.5%,due to the fine particle content. Many attempts have
been to generalize this upward concave profile into a “master profile, see for example Blight
(1994).He found that beach concavity is mainly caused by the particle sorting.
As an alternative to one tailings slurry discharge point, multiple discharges from a main
pipeline through short pipelines mounted perpendicular to the main pipeline (spigots) are
used. Spigoting in low to medium concentration tailings handling results normally in an
increased sorting effect with coarser particles building up close to the discharges along the
perimeter of the confined embankment, thus contributing to its function. Thus, spigoting
means an increased initial slope, however, the downstream value normally attains the low
slopes typical for one-point discharge.
The experience with spigoting giving a higher average slope for a given flow rate and solids
concentration when split up in several discharge points with lower flow rates expresses the
ability of a large flow rate to cause the flow channels to adapt a flatter slope thus bringing the
deposition farther out along the beach.
Generally with conventional handling at a constant discharge flow rate, the effect of hindered
particle settling through increased volumetric solids concentration also brings the deposition
farther out. For example, a slurry of narrowly graded particles of about 100 μm in average
size will be transported about 50% longer distance before settling out for an increase in
solids concentration from 10 to 30% ( CUR 1992 ).The effect on the average slope is here not
obvious because the resistance to flow may be higher for the 30%-mixture.
A discharged tailings slurry characterized by very small particles, for example clay, can be
considered to flow homogeneously, i.e. like a viscous fluid. No settling out of particles will
take place and an increase in the clay content increases the viscous behaviour and the flow
resistance, corresponding to a larger slope for a constant flow rate. Large flow rates for a
constant concentration enable the slurry to gradually spread further, thus influencing the
average slope.
6.2 Thickened tailings deposition

Williams (2001) suggest that two flow regimes exist on a tailings beach related to the flow of
highly-concentrated tailings slurries. A slurry stream in a narrow confined channel with
turbulent flow transports the tailings slurry to a localised broader area where the flow attains
sheet-like laminar flow where particles settle together, so-called zone-settling. This is
recognised at the main mechanisms for profile formation. Laminar conditions cannot persist
and with the local slope steep enough, turbulent flow will be maintained in the self-formed
channel. Cyclic repetition along the beach will in this way determine a limiting equilibrium
slope in a macroscopic sense, Pirouz et al. 2005 and Pirouz et al. (2007).
7.Discussion and conclusions

Recent approaches for slope evaluations , for example by Pirouz et al. (2007) and others, may
provide a means of predicting the beach slope angle with tailings slurries of arbitrary
rheological characterisation. The reported findings explain to some extent the difficulty with
laboratory scale flume tests, because of the problem to obtain turbulent flow conditions.
An alternative to a rheologically-based slope modelling approach is now considered and

limited to slope scaling predictions related to the turbulent settling behaviour of particles
with sizes of about 100 μm, De Groot et al. (1998)
The channellised tailings flow equilibrium slope concept discussed above with a net settling
out of tailings is here related to the slope requirement for a balance between sedimentation
and erosion along a beach formed by dredging material from parallel discharges from several
large pipelines, De Groot et al. (1988).They investigated slope angles for various narrowly
graded fine sand medium concentration slurries with particles down to about 60 μm., i.e.
two-component “settling” type of mixtures. They showed how the slope is dependent not only
on the particle size but also on the discharge flow rate, with steeper slopes for lower discharge
rates. They quantified the results from small- and large-scale flume experiments and field
observations in an empirical relationship (CUR 1992), which here expresses the slope, Sm,
obtained for modelled (small) flow rates, Qm, and larger prototype flow rates, Qp, as follows:
0.45
Sp § qm ·
¨ ¸ (2)
Sm ¨q ¸
© p ¹
where Sp is the prototype slope and qp and qm are defined as Qm/Bm and Qp/Bp, where Bm and
Bp are characteristic widths between discharges. The formulation in Eq.(2) is now adopted
assuming discharge in two parallel and closely located pipelines both in the model and
prototype scale, each with diameter Dm and Dp, respectively. The solids concentrations in
Figure 3 for which the slopes start to increase are related to the discharge from 0.025 m
pipelines from the pilot-scale thickening tests. This diameter is taken as the pipe diameter, Dm
in Eq.(2), with Dp of the order of 0.1-0.15m and values of Qm determined by the capacities.
S p | 0.25 S m (3)
The estimation expressed in Eq.(3),briefly used by Engman et al. (2004), indicates that
disposal slopes for the applications considered here should be about 2.5 to 4% because
observed slopes during the tests were 10 to 16%, dependent on the circulating pump was
operating or not ( Figure 3).
Robinsky (1999) presented a series of laboratory-scale flume deposition slope results

together with the corresponding particle size distributions. Comparisons with the observed
slopes of 10 to 20% here confirm roughly the results by Robinsky for similar size
distributions also giving field slopes of about 4%. However, he did not explicitly express any
scale-up criterion. Simms (2007) found that the 1:4 estimation in Eq.(3) was reasonable in a
comparison with a rheological approach for slope estimations and scaling. Evaluation of
flume data with field observations in the study by Pirouz et al. (2007) indicated a ratio close
to 1:2.
The indicatively quantifications discussed here on how small discharge rates tend to give
higher slopes demonstrate flexibility possibilities for the long term management of the tailings
storage facility. Spigoting, or simply a split of the pipeline flow into two streams just before
the discharge end would be feasible, principally giving two discharge points. However, there
are reasons to use only one pipeline to transport thickened slurry over long distances, because
the pressure (energy) requirement increases with a decreased pipeline diameter.
Conceptually with pastes, the deposited tailings has a planar slope, i.e. no concavity.
Experiences to date indicate that some degree of concavity occurs For example, the deposition
scheme at Kid Creek described above has a slope concavity of about 2.5 m on a radius of
1000m from a central discharge point. It is important to consider this effect in a long term
design perspective. Significant densification (consolidation) and deposit strength results often
from natural evaporation in warm areas and through freeze-thaw cycles in cold regions. For
example, the entire in situ deposit at the Kidd Creek disposal area described earlier is
consolidated to near the shrinkage limit at C about 56%,corresponding to a density of 2175
kg/m3. Simms (2007) related compaction effects to cyclic deposition schemes also giving
steeped slopes.
With the small amount of water reaching the disposal area with the highly thickened tailings
an increased potential for dust generation can be expected compared to conventional
deposition. However, it has been suggested that the lack of segregation with better packing
should make the surface more resistant to wind erosion, as indicatively demonstrated in a
laboratory-scale comparison by Robinsky (1999).The total surface area exposed to the wind
may also be limited, especially if a central discharge system is avoided.
The elevated location of the paste-thickener on the ridge close to the disposal area in Figure 1
are considered to be layered as paste in the vicinity of the thickener over several years and
then after about 20 years with pipeline lengths of up to 900m. During the first years of
operation only one thickener underflow pump is required. The reliability and performance for
various pump configurations and depositional schemes are planned to be investigated
systematically in order to meet the most effective long-term solution for paste thickener
operation, pipeline distribution and placement in the disposal area as the discharge point
advances over the life of the project.
8. References
ASCE Task Committee on Hydraulics, Bank Mechanics, and Modelling of River Width
Adjustment (1998) J. Hydr.Engrg.,Vol.124,Issue 9,pp903-917.
Blight G.E.(1994) The master profile for hydraulic fill tailings beaches, Proc. Inst. Civ Engrs
Geotech.Engng.,107,Jan. pp 27-40.
CUR (1992): Artificial sand fills in water, Centre for civil eng. research and codes, The
Netherlands, A.A. Balkema, Rotterdam
De Groot M.B., Heezen F.T., Mastbergen D .R. and Stefess H.(1988) Slopes and densities of
hydraulically placed sands, Hydraulic Fill structures, ASCE Geotechnical Division Speciality
Conference, Colorado State University, Ft. Collins, U.S.A.
Engman, M., A.Sellgren, Å. Sundqvist, I.Goldkuhl and T. Wennberg (2004):Users
perspective on the design of high density base metal tailings systems, Proceedings, 11th Int.
Conference on Tailings and Mine Waste’04,Vail, CO, U.S.A, October.
Fell R., Macgregor,P., and Stapleton D.(1992) Geotechnical engineering of embankment
dams, Balkema
Fourie, A. (2002): Important material characterisation of thickened tailings, Proceedings, Int.
seminar on paste and thickened tailings, Australian centre for geomechanics, May.
Hollow J.T. ,Lin H.K.,Walsh D.,E. and White D.M. (2007) Development of an energy
balance model to predict the economic impact of installing a tailings wash thickener at the
Fort Knox Mine, Fairbanks, Alaska.
Minerals and Metallurgical Processing,Vol.24,No.2,May ,pp 81-96.
Jewell, R. J., Fourie, A.B. (2006) Paste and thickened tailings –a guide. Australian Centre for
Geomechanics, University of Western Australia, Nedlands, Western Australia, 360p
Oxenford J. and Lord E.R. (2006) Canadian experience in the application of paste and
thickened tailings for surface disposal. Proceedings of the 7th International Seminar on Paste
and Thickened Tailings, April, Limerick, Ireland, pp 93-105.
Pirouz,B. ,Kavianpour, M. R. and Williams, P.(2005) Thickened tailings beach deposition-
field observations and full-scale flume testing. Proceedings of the 8th International Seminar
on Paste and Thickened Tailings, April, Santiago, Chile.
Pirouz, B.K. and Williams, P.(2007) Prediction of non-segregating thickening tailings beach
slopes-a new method. Proceedings of the 10th International Seminar on Paste and Thickened
Tailings, Perth, Australia, March, pp 315-328.
Pullum L.(2007) Pipelining tailings, pastes, and backfill. Prediction of non-segregating
thickening tailings beach slopes-a new method. Proceedings of the 10th International Seminar
on Paste and Thickened Tailings, Perth, Australia, March, pp 113-129.
Robinsky, E. I. (1999): Thickened Tailings Disposal in the Mining Industry. Quebecor
Printpak, Toronto, Canada
Simms P. (2007) On the relation between laboratory flume tests and deposition angles of high
density tailings. Proceedings of the 10th International Seminar on Paste and Thickened
Tailings,Perth,Australia,March,pp329-338.
Wennberg T. and Sellgren A. (2007) Pumping evaluations with paste tailings thickened close
to the surface disposal area ,Proceedings, The 17th Int. Conf. on the Hydraulic Transport of
Solids, Cape town, South Africa, May, pp 561-574.
Williams, M.P.A. (2001) Tailings beach slope forecasting-a review. High density and
thickened tailings conference, Pilanesberg, South Africa.
Winterwerp, J.C., de Groot, M. B., Mastbergen, D. R. and Verwoert H. (1990)
Hyperconcentrated sand-water mixture flows over a flat bed.JHydr.Eng.,Vol.116,pp36-54
Paper D
Feasibility considerations of high density disposal at the LKAB Svappavaara

concentrator
Wennberg T, Sellgren A and Sundqvist Å
Proceedings, The Int. Conference, Securing the Future 2005, June 27-30, Skellefteå, Sweden
Feasibility considerations of high density tailings disposal at the LKAB Svappavaara
concentrator.
Thord Wennberg1, Anders Sellgren2, Åsa Sundqvist1

1
Luossavaara-Kiirunavaara AB, Sweden
2
Luleå University of Technology, Sweden
Abstract
A pilot-scale deep-tank thickening test shows that the resulting minimum solids concentration
by mass giving a non-segregating mixture is about 70% corresponding to a volumetric
concentration of 44%. Flow related results from laboratory-scale flumes and pipelines
together with viscometric measurements and slope observations indicate that 90 tonnes of dry
tailings per hour can be disposed of in slopes of up to about 4%. With this, the time needed
for costly dam rises increases from 3 years to over 20 years. Preliminary economical
comparisons indicate that the deep tank thickener shall be located at the disposal area with
high density pumping of up to about 1 km to the central part. The costs for the three
considered alternatives may be rather similar, dependent on the actual pumping energy
required to transport the thickened tailings. It can be economically attractive to use centrifugal
pumps instead of positive displacement pumps. Various designs and arrangements of deep
tank thickeners also influence the pumping energy requirement and the possible slope. Small
discharge rates tend to give higher slopes. If the slope requirement becomes critical, or if
economically justified in the total system, a split of the pipeline flow into two streams just
before the discharge end would be feasible, principally giving two discharge points. The
further work is now directed to clarify the field-scale pumping energy requirement and slopes
for various thickener concepts together with environmental evaluations.
Introduction
Luossavaara-Kiirunavaara AB (LKAB), an iron ore company with mines in northern Sweden

is continuously considering new technologies for the handling, transportation and disposal of
waste rock and tailings. In the LKAB concentrator in Svappavaara about 0.7 Mtonnes of dry
tailings per year is transported nearly 1 km as a slurry in parallel pipelines connected to a
flume discharging into a tailings pond. In order to limit the cost of tailings management in the
existing area surface disposal of a highly thickened slurry at slopes of 3 to 5% has been
considered. The time needed for costly dam rises in the disposal area increases schematically
from 3 years with conventional handling with maximum slopes of about 1% to over 20 years
with thickened disposal.
The thickening may take place adjacent to the plant or close to the disposal area (1 to 2 km
away). The location of the thickener is mainly a balance of the cost for high pressure pumping
and the disadvantage of having the thickening facility remote from the rest of the processing
taking place in the concentrator.
The objective here is to show indicative flow related results from laboratory-scale flumes and
pipelines together with viscometric measurements and slope observations and to evaluate and
discuss the overall economic feasibility of various locations of the thickening. This study does
not include any direct environmental evaluations. The experimental work was carried out in
connection with pilot-scale deep-tank thickening tests at the Svappavaara concentrator by
LKAB and Outokumpu Mintec OY.
Thickener location
The pilot-scale thickening tests showed that the minimum solids concentration by mass giving
a non-segregating mixture that could be disposed of at slopes of 3 to 5%, was about 70%. The
corresponding volumetric solids concentration is 44%.The hourly capacity of dry tailings
considered is 90 tonnes, which is to be pumped in a 0.3 m-diameter pipeline to the deep-tank
thickener at a solids concentration by mass of about 10%.
After some consideration, it was decided to study two alternatives with location of the
thickener close to the disposal area, in addition to the alternative with the thickener located at
the concentrator. The alternatives denoted 1, 2 and 3 are shown schematically in Figure1.
C
+ 369 3
Highest point
+ 411
2
1 + 369
+ 342
W + 359
N
S
+ 365
Legend D
D: Disposal area
S: Settling pond
W: Water pond
Figure 1. Schematic sketch of the three alternative locations of the high-density thickener at
the Svappavaara concentrator.
In alternative1 in Figure 1 the location of the thickener is at the north-east corner of the
present disposal area about 1000 m from the concentrator. The discharge from the tailings
flume in the present handling system is close to the considered thickener location in
alternative 1. In alternative 2, the thickener is located on a plateau on the hill at the north side
of the disposal area. With a location of the thickener at the concentrator and with alternative
1, a ramp is needed to pump the thickened slurry in an elevated pipeline to the centre of the
disposal area. Pumping of thickened slurry is only required a short distance with the elevated
discharge from the hill in alternative 2, see Figure 2.
1000 m 1000 m
Concentrator Thickener +391

Water
+369
1:25
Pump
Pump Tailings
1800 m 300 m
Thickener
+411
Water Pump
Concentrator
+369
Pump
+365
Figure 2. Schematic description of the alternatives 1 and 2 in Figure 1 with location of the
thickener in the disposal area.
In Figure 2, the surplus water from the thickening is schematically circulated directly to the
concentrator. Today the solids concentration by mass in the tailings slurry from the
concentrator is about 5%, corresponding to 1710 m3/h of water. With high density thickening,
only about 39 m3/h water is discharged together with the tailings at the disposal area. This is
only a fraction (2.2%) of the water that is passing through the disposal area today. The 10%-
slurry (810 m3/h water) to the deep tank thickener is obtained by trimming the operation of
the existing conventional in-plant thickener. With high density thickening the remaining water
associated with the conventional handling, 900 m3/h, is considered to be pumped in existing
infrastructure to a rim (elevation +385m) close to the north-eastern part of the disposal area
and then transferred in a separate pipeline to the clarification pond for all three alternatives,
see Figure 1. In alternative 1, the surplus water from the deep tank thickening system, 810-
39= 771m3/h is added to the transfer line. In alternative 2, this water is transported by gravity
to the settling pond in a pipeline from the elevated thickener.
Experimental results
The initial measurements aimed at establishing particle size distributions that represent the
operating conditions of the concentrator. It was found that the thickening and handling system
must be robust and able to handle a span of particle size distributions due to process
variations. The tailings products had average particle sizes from 25 to 50 microns with 20 to
30% of less size than 10 microns. Top sizes were from about 400 to 500 microns. The solids
density was about 3000 kg/m3
Pilot-scale thickening test
The Outokumpu pilot-scale thickener is about 10 m high with an inner diameter of 1 m.

Tailings slurry is taken out for the test from the regular tailings pipeline system, see Figure 3.
The thickened slurry flows by gravity in an 11 m long plastic hose with inner diameter about
25 mm from the thickener to the deposition test on the floor.
3.
2.
+15 m
+14 m
Water
+10m
1. 4.
6. Slurry 5.
+5 m
7.
+0 m
1. Flow from existing tailings pipeline 5. Overflow water

2. Top-size control 6. Flocculating works
3. Mixing tank 7. Deposition test on the floor
4. Thickener
Figure 3. Schematic sketch of the pilot-scale test arrangement.
Deposition slopes
The resulting deposition slopes in Figure 3 were mainly 14 to 25% with concentrations in
excess of 70% by mass. The diameter of the cone-like deposition was about 6 m. A plexiglass
flume was also used for estimation of the deposition slope (angle of repose) through batch
experiments in an inclined test apparatus developed by Sofra and Boger (2001),see Figure 4.
The tailings slurry was stirred with a mixer and then placed in the flume reservoir upstream of
a movable gate.
Figure 4. Schematic sketch of the plexiglass flume for estimation of the deposition slope
In general the slurry flowed out of the flume at concentrations less than 65%. With values of
about 75% the slurry did not flow at all when the gate was removed. The results are shown in
Figure 5 together with slopes observed from the 25 mm-diameter pipeline discharge shown in
Figure 3.
30
F lo o r, F ig. 3
B a tc h , F ig. 4
25
20
Sl o pe (%
15
10
0
60 65 70 75 80
C o n c e n tr a tio n b y m a ss (% )
Figure 5. Observed deposition angles from 25 mm diameter discharge pipelines and batch
flume tests.
It follows from Figure 5 that average slopes of 10 to 15% were obtained from concentrations
by weight of 70 to 72%.
Slump test
The standard method for consistency measurements in the concrete industry has been
modified and adopted for yield stress estimations in connection with disposal of mine tailings,
simply by using a cylindrical tube, Pashias et al. (1996). Here a plastic tube was used with the
height, S, and diameter 75 mm. The tube is filled with tailings slurry and lifted where after the
settling (”slump”), z, is measured, see Figure 6.
Figure 6. Schematic sketch of a “slump” test where z is the slurry settling and S denotes the
height of the tube. After Pashias et al.(1996).
The results from the slump tests are mainly intended for comparisons of thickener
performance data, and results are available in Engman et al. (2004)
Viscometric tests
An energy balance relationship for the thickened slurry was used to roughly evaluate the
friction loss gradient in the 25 mm-diameter plastic hose schematically shown in Figure 3. In
this way, the hose served as a tube or extrusion viscometer. Indicative tests were also carried
out with a few samples with a vane-type viscometer, where resulting yield stresses varied
from about 90 to 135 Pa. High-density tailing slurries behave in a highly non-Newtonian way,
often exhibit a yield stress, meaning that they behave as a solid until sufficient force is
applied. The yield stress is defined as the minimum stress required causing the solid-liquid
mixture to flow.
The 25 mm-hose and results here and 50 mm-pipeline and continuous flume flow results for a
similar tailings product by Engman et al. (2004) are presented versus the rheological scaling
parameter 8V/D in Figure 7, where V is velocity and D the pipeline diameter.
1000
Pipeline
Flume
100
Shear stress (Pa)
10
1
0 1 10 100 1000
8V/D (s-1)
Figure 7. Shear stresses versus the rheological scaling parameter 8V/D from pipeline and
flume evaluations. From Engman et al.(2004).
The results presented in Figure 7 at low values of the scaling parameter indicate yield stresses
in the same range as found with the vane-type viscometer. In the pipeline transport application
8V/D-values of 100 to 200 may be considered. Indicatively, resulting shear stresses from
Figure 7 should be in the range of 100 to 200 Pa. The pipe wall shear stress, τw, is related to
the pipeline friction loss gradient, i , expressed in metres of slurry per metre of pipe, and the
required pressure gradient, Δp/Δx, through the following relationship where ρ is the slurry
density:
('p / 'x) D UgiD

Ww (1)
4 4
With the capacity 90 tonnes/h, the flow rate of slurry is 68 m3/h corresponding to an operating
velocity of about 2.2 m/s in a 0.1 m-diameter pipeline, which is preliminary considered here.
The 8V/D-value is then 175 indicating a shear stress between 100 to 200 Pa in Figure 7.
Recent work with slurries of the type considered here has shown that they may not behave in
a completely non-settling way,i.e. flowing homogeneously during pipeline transportation, see
for example Pullum (2003). This may influence the accuracy of the rheology-oriented pipeline
friction loss evaluations used here. In addition, the indicative nature of the small-scale results
in Figure 7 make it reasonable to consider both 100 and 200 Pa. With these values in eq.1,
then the pressure requirement may vary from 4 to 8 MPa/ km.
Cost comparisons-Discussion
The maximum power requirement to overcome pipeline friction for pumping the thickened
slurry in alternative 1 is then 8 MPa and 16 MPa for alternative 3. The transportation of
highly thickened slurry with working pressures in this range is normally carried out with
positive displacement type pumps. The pilot-scale thickening test results indicate a thickener
with a height of about 10 m and a diameter that is more than twice the height. The preliminary
evaluated total investments and slurry pumping costs are summarized in Table 1.
Table 1.Overall preliminary investments including design, overhead and start up and direct
slurry pumping energy costs for the three alternatives shown in Figures 1 and 2. Electrical
energy =0.25 SEK/kWh (1 SEK= 0.11 EUR= 0.15 US$, January 2005)
Alternative (Figure 1) 1 2 3
Investments (MSEK) 44 49 69
Pumping energy costs (MSEK)
Thickener feed lurry 0.3 0.7 0.1

Thickened slurry 0.4 0.1 0.8
Total slurry pumping energy costs 0.7 0.8 0.9
The total operating costs including thickening, regular maintenance, additional energy,
pipeline and pump wear etc., are roughly two times the total slurry pumping energy costs in
Table1. It is assumed in alternative 1 that the deep tank thickener can be located in a way that
overflow water flows by gravity with the excess water from the plant to the water pond or
directly back to the plant, see Figures 1 and 2.
The investments for pumps and pipelines to transport the low and high density slurries
account for about 50% of the total capital costs. The thickened slurry pumping costs in Table
1 are directly related to the high density pressure requirement. Assume that a representative
shear stress here is 100 instead of 200 Pa, which means that the pressures would be 4
MPa/km, corresponding to 4 MPa in alternative 1 and 8 MPa in alternative 3.The total slurry
pumping costs in alternative 1, 2 and 3 in Table1 then correspond to 0.5, 0.75 and 0.5 MSEK,
respectively. The decreased pressure requirement lowers the investment costs only to some
extent, because the costs of positive displacement pumps are relatively insensitive to
variations in this pressure range.
Centrifugal pumps in series may be an economically interesting alternative, particularly with

pressures of about 4 MPa/km. Centrifugal pumps are used for thickener underflow circulation
and feeding to positive displacement pumps in high-pressure applications or to pump
thickener underflow slurry short distances. Provided both systems are technically feasible,
using a positive displacement pump instead of centrifugal pumps in series generally implies
capital costs that are ten times greater, Cowper (1999). Loop tests with centrifugal pumps and
yield-like slurries with pipeline shear stresses of 100 to 200 Pa have shown that these pumps
may be feasible, see for example Sellgren et al. (2005).
About seven centrifugal pumps in series are required based on a pressure of 4 MPa per km to
overcome pipeline friction. The energy cost will be about 1.7 times higher than for positive
displacement pumps because the efficiency of a positive displacement pump may be slightly
over 80% compared to nearly 50% for a centrifugal pump in this application. However, the
much lower capital cost makes the centrifugal pump alternative attractive. With this
alternative, the total pumping energy costs in Table 1 are 0.65, 0.7 and 0.8 MSEK for
alternative 1, 2 and 3, respectively, corresponding to investments of 40, 49 and 50 MSEK.
Deposition slopes
Deep-tank thickeners are sometimes equipped with a circulating centrifugal pump
arrangement in order to ensure consistent discharge. It is generally experienced that this
agitated slurry flows more easily in pipelines and gives a lower slope than for non-agitated
slurries. This was observed in connection with pilot-scale thickening tests including a
circulating pump, Engman et al.(2004).
The time needed for costly dam rises in the Svappavaara disposal area increases from 3 years
with conventional handling to over 20 years with thickened disposal at a slope of 4%. The
small-scale tests carried out here indicate field slopes of 3 to 5%, Engman et al.(2004). From
their analysis and literature references it follows that small discharge rates tend to give higher
slopes. Therefore, if the slope requirement is critical, a split of the pipeline flow into two
streams just before the discharge end would be feasible, principally giving two discharge
points. However, there are reasons to use only one pipeline to transport thickened slurry over
long distances, because the pressure (energy) requirement increases with a decreased pipeline
diameter.
Conclusions
With a solids concentration by mass of about 5% from the processing plant and with about
10% going to the deep-tank thickening, about 900 m3/h of water will be circulated via the
clarification pond to a reservoir at the plant. The annual energy cost for this is of the same
order of magnitude as the high pressure (16 MPa) thickened tailings pumping about 2 km in
alternative 3. The goal is to reach a water balance situation at the plant where circulation of
large quantities of water through the water pond can be avoided.
Preliminary economical comparisons indicate that the deep tank thickener shall be located at
the disposal area with high density pumping up to about 1 km to the central part. The costs for
the three considered alternatives may be rather similar, dependent on the actual pumping
energy required to transport the thickened tailings. It can be economically attractive to use
centrifugal pumps instead of positive displacement pumps.
Various designs and arrangements of deep tank thickeners also influence the pumping energy
requirement and the possible slope. Small discharge rates tend to give higher slopes. If the
slope requirement becomes critical, or if economically justified in the total system, a split of
the pipeline flow into two streams just before the discharge end would be feasible, principally
giving two discharge points. The further work is now directed to clarify the field-scale
pumping energy requirement and slopes for various thickener concepts together with
environmental and water balance evaluations.
Acknowledgements
The results presented and discussed here are based on measurements, observations and
analyses carried out solely by the authors and personal at LKAB together with Monica
Engman, presently with Sydkraft Vattenkraft AB. Outokumpu Mintec OY has reported their
pilot-scale thickening results and conclusions in confidential reports to LKAB.
References
Cowper N.T, 1999: Rheology, a key factor in pumping and slope disposal of high density
mine tailings Proceeding, Rheology in the mineral industry II. E.J. Wasp. Ed., Kahudu,
Oahu, Hawaii, U.S.A.
Engman, M., A.Sellgren, Å. Sundqvist, I.Goldkuhl and T. Wennberg ,2004: Users perspective
on the design of high density base metal tailings systems, Proceedings of the11th Int.
Conference on Tailings and Mine Waste’04, Vail, CO, U.S.A, October, A.A. Balkema
Rotterdam.
Pashias, N., Boger D.V., Summers J. and Glenister, D.J. ,1996: A Fifty Cent Rheometer for
Yield Stress Measurement. Society of Rheology, Inc., Journal of rheology, 40(6), pp 1179-
1189.
Pullum, L.,2003: Pipeline performance, Proceedings, Int. seminar on paste and thickened
tailings, Australian centre for Geomechanics, Melbourne, May.
Sellgren A., Addie G.R., Whitlock L., 2005: Technical-economical feasibility of using
centrifugal pumps in high-density thickened tailings slurry systems, paper to be presented
at: Paste 2005,Int’l seminar on paste and thickened tailings, Santiago, Chile, April 20-22.
Sofra F., Boger, D.V.2001: Slope prediction for thickened tailings and pastes. Proceedings of
the Eighth International Conference on Tailings and Mine Waste’01, Fort Collins,
Colorado, 16-19 January.A.A. Balkema Rotterdam, 75–83.
Paper E
Users perspective on the design of high density base metal tailings systems
Engman M, A Sellgren, Å Sundqvist, I Goldkuhl and T Wennberg
Proceedings, 11th Int. Conference on Tailings and Mine Waste’04,Vail, CO, U.S.A, October
Proceedings, Tailings and Mine Waste’04,Taylor & Francis Group, London, ISBN 041535939 2
Users perspective on the design of high density base metal tailings

handling systems
M. Engman
Loussavaara-Kirunavaara AB, presently with Sydkraft Vattenkraft AB, Sweden
A. Sellgren
Lulea University of Technology, Sweden
Å. Sundqvist & T. Wennberg

Loussavaara-Kirunavaara AB, Sweden
I. Goldkuhl
MultiServ AB, Sweden
ABSTRACT: Various aspects of transportation and disposal of tailings at high solids concentrations are
presented and discussed based on measurement and evaluations of flow-related parameters in laboratory-scale
flumes and pipelines together with viscometric data and deposition slope observations. The work was carried
out in connection with pilot-scale deep tank thickening tests at two base metal concentrators in central and
northern Sweden. The resulting minimum solids concentration by mass giving a non-segregating mixture was
about 70% corresponding to a volumetric concentration of about 44%. The tailings products had average
particle sizes from 25 to 50 microns with 20 to 30% of less size than 10 microns. Both users need some
flexibility with respect to the size distribution due to processing variations. The thickening is preliminary
considered to be in the disposal area or between the concentrator and the deposition. The results indicate that
disposal slopes in the applications should be about 4% based on maximum 16% for the smaller flow rates in
the test. The flow rate dependence means that a split of the pipeline flow into two streams would give higher
slopes. However, the split should be close to the discharge in order to avoid the disadvantage of having two
smaller pipelines in parallel which increases the pumping energy requirement. Scaling criteria for field slope
evaluations and effects of pump passages and circulatory arrangements need to be clarified.
order to obtain the properties that give virtually no

1. INTRODUCTION drainage of water and segregation of particles.
The concept of a high degree of thickening means
There is an increased interest throughout the world that the deeply thickened tailings slurry is
in the treatment and processing of tailings from discharged from an elevated or man made position.
mineral beneficiation plants in order to reduce the With the higher angle more volumes can be filled
volumes and sizes of tailing dams. The reasons are per unit surface area for a constant dam height. With
generally a combination of environmental and AMD- generating tailings multi-layer capping may
economical considerations. The economical be needed. The reduced cost for dams, capping and
feasibility of a high degree of tailings thickening is return water pumping can be considerable over the
determined by the long-term impact in the disposal life of the mine. A smaller area also concerns dam
area. safety. Some reported tailings dam failures and
A tailings mixture that is produced and handled incidents may be related to large amounts of water in
with an extremely high solids concentration is often the disposal area. The local conditions strongly
called paste, a term that is associated with bauxite determine the feasibility and economical
processing and backfilling in underground mines effectiveness of different degrees of thickening. An
with use of binder to obtain high strength. The important factor is the location of the thickener;
amount of particles smaller than 20 m normally pipeline pumping over long distances requires very
exceeds 15 % and with an average particle size that high working pressures.
may vary from 20 to 100 m. The solids
concentration by volume must often exceed 45 % in
2. OBJECTIVES AND SCOPES with respect to the particle size distribution, the span
is given in Figure 1.
The cases considered here concern surface disposal In order to get the high solids concentration for
slopes of 3 to 5% in order to limit the cost of tailings paste, tank thickeners are often combined with
management in existing disposal areas. In the mechanical (filter) dewatering. Only deep tank
Loussavaara-Kirunavaara AB’s (LKAB’s) thickeners are considered here because no distinct
Svappavaara iron ore concentrator in northern slope requirement is set for the disposal. The study
Sweden about 90 tonnes/h of dry tailings are was carried out in connection with pilot-scale
considered to be thickened at the plant or in the thickening tests by Outokumpu Mintec OY at LKAB
disposal area about 3 km away. In MultiServ AB’s and Dorr-Oliver Eimco at MultiServ AB.
plant in central Sweden for recovery of metal from
slag about 45 tonnes/h are planned to be thickened in
the disposal area about 500 m away from the
concentrator. Both operations need some flexibility
100
90
80
70
Percent passing
60
50
40
30
20
10
0
1 10 100 1000
Particle size ( m)
Figure 1.The particle size distribution span for the tailings products considered in this study. Solids density 3000 kg/m3.
The objective is here to present experimental

results and to discuss various aspects of thickening,
transportation and disposal of tailings products 3. PILOT SCALE THICKENING SET-UP
within the span in Figure 1. Some of the results have
also been presented in a recent paper, Engman et al. The Outokumpu pilot-scale thickener is about 10 m
(2004). Detailed results will here be focused on high with an inner diameter of 1 m. Tailings slurry at
flume flow and deposition slope scaling. Cw 3-4% is taken out for the test from the regular
For practical reason, the solids content in the tailings pipeline system, see Figure 2.
tailings slurry is related to the concentration by mass The thickened slurry flows by gravity in an 11 m
(weight), Cw. A Cw-value of 70% corresponds to a long plastic hose with inner diameter about 25 mm
concentration by volume and slurry density of 44% from the thickener to the deposition test on the floor.
and 1880 kg/m3, respectively. The pilot-scale In a similar arrangement at MultiServ AB the
thickening results showed that Cw about 70% is a Dorr-Oliver Eimco pilot unit was fed via an agitated
minimum for a mixture where solids do not tank from a sampler in the main tailings line where
essentially settle or segregate. Cw was 30 to 35%. Dilution water was added to the
tank to achieve a suitable lower solids concentration.
The unit was about 5 m height with a diameter of 25 mm diameter plastic hose. A 0.05 by 0.038 m
about 1.5 m, see Figure 3. centrifugal pump with impeller diameter 0.225 m
The Eimco pilot-scale deep cone thickener is can be used to circulate slurry in an about 6 m long
equipped with a circulating pump arrangement loop with inner pipeline diameter 0.05 m. The pump
which in full scale applications ensures consistent is instrumented for rpm, voltage and ampere
discharge. Circulated slurry pass through under readings and a magnetic flow meter is installed in
mixing in the lower portion where thickened slurry the loop.
is taken out with a peristaltic pump connected to a
3.
2.
+15 m
+14 m
Water
+10m
1. 4.
6. Slurry 5.
+5 m
7.
+0 m
1. Flow from existing tailings pipeline 5. Overflow water

2. Top-size control 6. Flocculating works
3. Mixing tank 7. Deposition test on the floor
4. Thickener
Figure 2. Schematic sketch of the pilot-scale test arrangement at the LKAB plant
In order to roughly evaluate pipeline friction modified and a pressure gauge was installed when
losses and pump characteristics the loop was the system was erected at MultiServ AB.
Figure 3. Schematic sketch of the Dorr-Oliver Eimco pilot thickener set-up with a circulating pump.
4. LABORATORY-SCALE EXPERIMENTS
4.1. Flumes
A plexiglass flume was constructed at LKAB for
estimation of the deposition slope (angle of repose)
through batch experiments in an inclined test
apparatus developed by Sofra & Boger (2001),
Figure 4.
The tailings slurry was stirred with a mixer and
then placed in the flume reservoir upstream of a
movable gate. The height, H, of the slurry was
established. When the gate is lifted the slurry flows
out into the flume which here was horizontal (θ =0).
The heights and the length of flow were measured
when the slurry came to rest, see Figure 5.
Figure 4. Schematic sketch of the plexiglass flume

Figure 5. Recorded heights and extension of flow (L) after an experiment in the flume in Figure 4. After Sofra & Boger (2001)
The angle of repose, r, was thereby calculated where

from: ' y
H H2 y (3)
ttan 1 ( 1 ) (1) gS
r
L and:
A similar-sized flume was manufactured in steel z
for robust batch experiments and continuous flume s' (4)
S
flow observations at MultiServ AB. Equations 2, 3 and 4 then gives:
y gS 0.5(1 z/S) (5)
4.2. Slump tests
The standard method for consistency measurements
4.3. Pipeline and viscometric tests
in the concrete industry has been modified and
adopted for yield stress estimations in connection With flow rate measurements, pump operating data,
with disposal of mine tailings, simply by using a pressure and tank elevation readings in the
cylindrical tube, Pashias and Boger (1996). Here a MultiServ AB’s set-up (see Figure 3), it is possible
plastic tube was used with the height, S, and to roughly evaluate pipeline friction losses and pump
diameter 75 mm. The tube is filled with tailings characteristics through application of energy balance
slurry and lifted where after the settling (”slump”), relationships for the pumped slurry. A similar
z, is measured, see Figure 6. balance was used to evaluate the friction loss
gradient in the 25 mm diameter plastic pipe
schematically shown in Figure 2 in Svappavaara.
Viscometric tests were carried out with a few
samples with a vane-type rheometer available at the
LKAB R and D centre in Malmberget.
5. RESULTS
Figure 6. Schematic sketch of a “slump” test where z is the 5.1. Deposition slopes
slurry settling and S denotes the height of the tube. After
Pashias and Boger (1996). The deposition tests on the floor at the Svappavaara
concentrator created cones with diameters of 7 to 8
m and heights of 0.5 to 1 m. For Cw-values of 70 to
Based on experiments and dimensional analysis, 72%, the average slopes were 14 to18%,
Pashias and Boger expressed a dimensionless yield respectively. The deposited slurry had not been
stress, τy’, in the following way: agitated through any pump passage before disposal
'
0.5 0.5 s ' (2) on the floor, Figure 2. The effect of agitation on the
y
rheology was investigated with a mixer. Any
obvious influence on the viscous behaviour in slump 30
and viscometric measurements was not found. Floor, Fig. 2
During the process of start-up at MultiServ AB Batch
and reaching optimum performance of the thickener, 25 Trench, Fig.7
underflow was discharged from the 25 mm-pipeline Trench, circulated
or from the steel flume into a dug out hole with an Batch, circulated
20
initial slope of about 1:1 for the first 4 m.
For deposition observations during optimum
Slope (%)
steady state operation of the thickener a trench with 15
a trapezoid section was dug out in the existing
tailings dam, Figure 7.
10
The ground slope in the trench was 0.4% for the
initial 10 m length used in the test. With slurry
deposited during about 25 h from the underflow 5
pump a slope of nearly 16% was attained for a
length of 6 m. Continuing trials were carried out
0
during about 15 h with agitated slurry placed on the 65 70 75
earlier deposited material. The slurry was agitated Cw (wt-%)
through circulation with the centrifugal pump. It was
observed that this slurry flowed more easily down
the slope. The slope for this agitated slurry was
Figure 8. Observed deposition angles from 25 mm diameter
about 10%, which was less than for unagitated discharge pipelines and batch flume tests. Unfilled and filled
slurry. Limited slump measurements indicated only marks correspond to the tests at LKAB and MultiServ AB,
a slight decrease in yield stress. respectively.
5.2. Indicative pipe loop results

The circulation system in the thickener at MultiServ
AB was used only in the end of the test period in
connection with over 50 h of continuous operation.
In order to start the circulation system flushed water
Figure 7. Section of the trench used for deposition observations had to be introduced in the loop and the initial
at MultiServ AB
operation took place in a “sluggish” manner with
several stops with repeated flushing periods.
About 15 measurements were made of flow rate,
Batch flume tests were carried out with both the
pressure and pumps operating data. There was a
plexiglass and the steel flumes and evaluated with
trend of a circulatory effect with reduced pressure
equation 1. Generally with Cw-values less than 65%
readings with time. Evaluated flow rates and
the slurry flowed out of the flume. With values at
corresponding pumping heads were comparatively
about 75% the slurry did not flow at all when the
low and outside the normal operating domain for the
gate was removed. The results are shown in Figure 8
pump.
together with slopes from 25 mm-diameter pipelines.
The friction losses gradient in the 0.05 m
It follows from Figure 8 that average slopes of 10
diameter steel pipe and the pump head was
to 15% were obtained from concentrations by weight
estimated from energy-balance relationships as
of 70 to 72% and that pump circulation in the
explained in 4.3. Elbow losses were simply related
thickener at MultiServ AB tended to lower the slope,
to water flow conditions.
thus affecting the behaviour of the slurry.
Of the 15 runs only two were considered to be
reliable with a reasonable matching of pipeline
system and pump operating data. It was not possible
to evaluate the effect of slurry on the pump head and
efficiency; the flow rates were less than 40 % of the
flow rates in the best efficiency domain. The
resulting pipeline data are shown in Table 1.
Table 1. Evaluated friction losses and operating conditions in Vd2 Vu2
the circulating loop system in Figure 3. ( yd ) ( yu )
Friction loss Shear stress Velocity 8V 2g 2g
L (7)
gradient w V D Sb S f
(m slurry/m) Pa m/s s-1 where yd and yu are down flow and up flow slurry
0.79 184 1.27 203 depths over the length ΔL. Vd and Vu are the
0.65 150 1.13 181 corresponding velocities and α is a correction factor
for the velocity distribution. Sb is the flume slope
and Sf is the average friction loss gradient, here
The shear stress, τ w, in Table 1 is related to the expressed in the following way:
friction loss gradient, i, and the pressure gradient, wP
Δp/Δx, through the following relationship: Sf (8)
gA
( p / x) D gi
giD
w (6) where τw is the mean wall shear stress, P is the
4 4 average wetted perimeter and A is the average
where ρ and D is the slurry density and pipeline section area. Eq. 7 is known as the direct step
diameter, respectively. formulation in textbooks in open channel water
flow.
5.3. Flume flow With flow rates of 10 -4 m3/s corresponding to
velocities of less than about 0.01 m/s, the velocity
Observations of continuous flume flow at Cw about heads in eq. (7) are negligible compared to other
70% and 11.5% flume slope showed that the flow terms. The average shear stress from equations (7)
had an oscillating pattern. Observed depths along the and (8) is then:
flume shown here are rough values indicating the
y yd gA
span. The flume needed to be inclined more than 10 w ( u Sb ) (9)
% in order to avoid overflow of slurry along the L P
sides. The slope was gradually increased up to about
27 % where approximately uniform conditions with
a nearly constant depth were observed.
The successively change in depth along the flow
in the flume corresponds to an average energy and
momentum balance which can be expressed in the
following way in an interval, ΔL:
Table 2. Calculated average shear stress and profile data for two different tests compared to observed average depths and an
approximate flow rate of 10-4 m3/s with Cw=70%. Slope = 11.5 %. Upper and lower values may represent the oscillating span.
Length along flume m 0 0.25 0.50 0,70 0,90
Observed depth m 0.1 - 0.090 0.080 0.060

0.1 - 0.055 - 0.040
Calculated depth m 0.097 0.090 0.080 0.060
0.078 0.056 0.045 0.040
Average τw Pa 117 128 141 164
171 151 106 77
Average velocity m/s 0.0051 0.0054 0.0059 0.0073
0.0057 0.0077 0.0100 0.0118
0.12
Observed depths
0.1
Calculated profiles
0.08
Slurry depth (m)
0.06
0.04
0.02
0
0.00 0.25 0.50 0.70 0.90
Length along flume (m)
Figure 9. Slurry depths versus flume length with data from Table 2
Observed depths have been matched to a slurry in Figure 11 represent average conditions from a few
profile through eqs. 7 to 9.The corresponding shear samples from both test locations. The variation in
stresses and profiles for ΔL-values of 0.25 and 0.20 estimated yield stresses varied from about 90 to 135
m are shown in Table 2 and Figure 9, respectively. Pa.
The oscillation pattern is shown in Figure 9 and The results presented in Figure 11 correspond to
the corresponding span in shear stresses can be seen an approximate yield stress of 110 Pa, and a low
in Table 2. increase of shear stress with increased shear rates.
Nearly uniform conditions with a depth of about Comparison with Figure 10 indicates yield stresses
0.045 m along the 0.9 m reach were obtained at 27 in the same range.
% slope, for an approximate flow rate of 0.9·10 -4 The pipeline shear stresses in Table 1, the
m3/s and Cw about 70%. The corresponding τ w and Svappavaara 25 mm-hose evaluation (Figure 2) and
velocity were 170 Pa and 0.01 m/s, respectively. the flume shear stresses in Table 2 and for uniform
flume flow conditions are presented versus 8V/D in
Figure 12 where D for the flume results is expressed
5.4. Slump and viscometric tests
as the hydraulic diameter or equivalent pipe
The results from the slump tests are presented in diameter.
Figure 10, and show evaluated yield stresses from
eq.5 at different solids concentrations. The rheogram
700
600
LKAB
MultiServ AB
500
400
(Pa)y
300
200
100
0
60 65 70 75 80
Cw (weight-%)
Figure 10. Results from slump tests. The inaccuracy may be related to variations in the particle size distribution and inconsistent
solids concentration measurements
200
Shear stress, (Pa)
100
0
0 100 200
-1
Shear rate (s )
Figure 11. Average rheogram for Cw-values from 70-to72%, representing both plant locations measured with a vane-type Bohlin-
viscometer.
1000
Pipeline
Flume
100
Shear stress (Pa)
10
1
0 1 10 100 1000
8V/D (s-1)
Figure 12. Shear stresses versus the rheological scaling parameter 8V/D from pipeline and flume evaluations.
6. DISCUSSION AND CONCLUSIONS for example Sellgren et al. (2002). Furthermore,

experimental work with a modified pump showed
that shear stresses in excess of 200 Pa had a
6.1. Pipeline pumping
moderate influence on the pump performance
The resulting shear stress versus the rheological (Whitlock and Sellgren, 2002). It is therefore of
scaling parameter 8V/D in Figure 12 indicates that interest to technically and economically evaluate the
an overall design shear stress of about 200 Pa may intermediate area where both centrifugal and
be used here. With the considered capacities 45 and positive displacement pumps are feasible.
90 tonnes/h, the flow rates are 34 and 68 m3/h,
respectively, in the MultiServ AB and LKAB
applications. The corresponding operating velocities 6.2. Deposition slopes
are 1.2 and 1.5 m/s in 0.1 and 0.125 m diameter The time needed for costly dam rises in the
pipelines, respectively. The pressure requirements Svappavaara disposal area increases from 3 years
(eq.5) are then 8 and 6.4 kPa/m, respectively. with conventional handling to over 20 years with
The location of the thickener is mainly a balance thickened disposal at a slope of 4%. Furthermore,
of the cost for high pressure pumping and the circulation of large quantities of water through the
disadvantage of having the thickening facility disposal area can be avoided. Principally similar
remote from the rest of the processing at the advantages can be seen at MultiServ AB.
concentrator. At MultiServ AB the location of the The solids concentrations in Figure 8 for which
thickener will be in the disposal area with underflow the slopes start to increase are related to batch
transportation lengths of about 100 m, corresponding experiments in a 0.2 m wide and 2 m long flume or
to a pressure of about 0.8 MPa. Assuming a location the discharge from 0.025 m pipelines from pilot-
of the thickener at Svappavaara between the scale thickening tests. Robinsky (1999) presented a
concentrator and the disposal area then the series of laboratory-scale deposition slope results
transportation distance would be about 1000 m together with the corresponding particle size
which corresponds to about 6.4 MPa in pressure distributions. Comparisons with the observed slopes
requirement. The transportation of highly thickened of 10 to 20% here confirm roughly the results by
slurry is normally carried out with positive Robinsky for similar size distributions also giving
displacement type of pumps. However, centrifugal field slopes of about 4%. However, he did not
pumps are used for thickener underflow circulation explicitly express any scale-up criterion. Small- and
and feeding to positive displacement pumps in high- large-scale flume experiments and field observations
pressure applications or to pump underflow slurry for fine sands by (CUR, 1992) indicate that the
short distances. slope, Sm, obtained for modelled (small) flow rates,
Loop tests with centrifugal pumps and yield-like Qm, and larger prototype flow rates, Qp, are related
slurries with pipeline shear stresses of 100 to 200 Pa as:
have shown that these pumps may be feasible, see
Sp
0
0..45 out solely by the authors, with exception for the
qm
(10) pipeline loop tests at MultiServ AB which were
Sm qp done in cooperation with personal from Dorr-Oliver
where Sp is the prototype slope and q p and qm are Eimco. This company and Outokumpu Mintec OY
defined as Qm/Bm and Qp/Bp, where Bm and Bp are have reported their pilot-scale thickening results and
characteristic widths (CUR, 1992) here simply taken conclusions in confidential reports to MultiServ AB
as the corresponding pipe diameters, Dm and Dp. For and LKAB, respectively. The authors gratefully
example, at MultiServ AB, Qp equals 0.0001 and Qm acknowledge the cooperative work mentioned above
0.009 m3/s with Dm 0.025 and Dp 0.1 m, and the pleasant time together during the test periods
respectively, then qp = 0.004 and qm = 0.09 m2 /s. with the involved personal from Outokumpu Mintec
These values inserted in eq. (10) means that: OY and Dorr-Oliver Eimco.
0.45 A special acknowledgement to Sydkraft
0.004
Sp Sm 0.25S m (11) Vattenkraft AB for the support of the first author,
0.09 Monica Engman, to fulfil this documentation as a
The indication is thus that disposal slopes should part of her graduate studies.
be about 2.5 to 4% in an application with a 0.1 m
diameter pipeline operating at 1.2 m/s, because
observed slopes during the tests were 10 to 16%, 8. REFERENCES
dependent on the circulating pump was operating or
not. Boger, D.V. & Sofra F. (2001): Slope prediction for thickened
As indicated in eq. (10), small discharge rates tailings and pastes. Proceedings of the Eighth International
Conference on Tailings and Mine Waste’01, Fort Collins,
tend to give higher slopes. Therefore, a split of the Colorado, 16-19 January.A.A. Balkema Rotterdam, 75–83.
pipeline flow into two streams just before the CUR (1992): Artificial sand fills in water, Centre for civil eng.
discharge end would be feasible, principally giving research and codes, The Netherlands, A.A. Balkema,
two discharge points. However, there are reasons to Rotterdam
use only one pipeline to transport thickened slurry Engman, M., Sellgren, A., Sundqvist, A., Goldkuhl, I. and
Wennberg, T. (2004): Some design considerations of high
over long distances, because the pressure (energy) density base metal tailings handling systems, accepted for
requirement increases with a decreased pipeline presentation at the 12 Int. Conference on transport and
diameter. sedimentation of solids particles, Sept 20-24,Prague, Czech
Scaling criteria for field slope evaluations need Republic
to be studied further, a need also expressed in a Fourie, A. (2002): Important material characterisation of
thickened tailings, Proceedings, Int.seminar on paste and
recent flume test study, Fourie and Gawu (2004). thickened tailings, Australian centre for geomechanics,
Furthermore, effects of pump passages and Melbourne, May
circulatory arrangements should be clarified. Most Fourie, A. and Gawu, S (2004): Post-depositional behaviour of
experiences so far of surface deposition and tailings high density tailings, Proceedings, Int. seminar of paste and
handling at very high solids concentrations include thickened tailings: Paste 2004,Capetown, South Africa.
Pashias, N. & Boger D.V. (1996): A Fifty Cent Rheometer for
clay minerals. For ground rock base metal mine Yield Stress Measurement. Society of Rheology, Inc.,
tailings, the finer (clay-sized) particles are not really Journal of rheology, November/ December
clay minerals but rock flour, Fourie (2002), a Pullum, L. (2002): Pipeline performance. Proceedings, Int.
dissimilarity also marked by Robinsky (1999). seminar on paste and thickened tailings, Australian centre
Recent work with slurries of the type considered for Geomechanics, Melbourne, May
Robinsky, E. I. (1999): Thickened Tailings Disposal in the
here has shown that they may not behave in a Mining Industry. Quebecor Printpak, Toronto, Canada
completely non-settling way flowing Sellgren, A., Addie, G. & Whitlock L. (2002): Using
homogeneously during pipeline transportation, see centrifugal pumps for highly concentrated tailings slurries,
for example Pullum (2002). This influence the Hydrotransport 15, Int. conference on the hydraulic
accuracy of the rheology-oriented pipeline friction transport of solids in pipes, Banff, Canada, June 3-5,669-
679
loss evaluations used here. These factors should also Whitlock, L., & Sellgren, A. (2002): Pumping paste with a
be clarified in more details. modified centrifugal pump, Proceedings, Tailings and mine
waste 02, Ft.Collins, Colorado, U.S.A., A.A. Balkema
Rotterdam, 155-163.
7. ACKNOWLEDGEMENTS
The results presented and discussed here are based

on measurements, observations and analyses carried

Transporting Highly Concentrated Slurries With Centrifugal Pumps

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Transporting Highly Concentrated Slurries With Centrifugal Pumps

Uploaded by

Copyright:

Available Formats

L I C E N T I AT E T H E S I S

ISSN: 1402-1757 ISBN 978-91-7439-169-5

Luleå University of Technology

1.1 Conventional and thickened tailings management

3.2 Conduit flow resistance

7KHSUHVVXUHJUDGLHQW p / x 3DP IRUWKHIORZLQDKRUL]RQWDOSLSHOLQHLV

3.3 Free surface flow (Paper E)

Pump head curve

Pump head curve, water

Efficiency curve, water

5.1 Flow resistance in pipelines and flumes (Paper B and E)

100 Pipeline diameter

5.3 Pump performance (Paper A and B)

5.4 Overall result summary

6.1 Centrifugal pump performance (Paper A and B)

50 100 150 200 250 300

6.2 Depositional conditions (Paper C and E)

Predicting the performance of centrifugal pumps when handling complex slurries

Wennberg T, Sellgren A and Whitlock L

Proceedings, 14th International Conference on Transport & Sedimentation of Solids particles,

Thord Wennberg1, Anders Sellgren2, Lee Whitlock3

KEY WORDS: complex slurries, pump performance derating, entrance losses

The performance of centrifugal pumps is influenced by the solids in the slurry

H Head ratio: HR = H/Ho

Abritrary HR or RH values from Fig. 2 are then obtained through concurrent

ANSI/HI (2004) provides a generalized derating method of the water performance

Apparent viscosity is the viscosity at a particular rate of shear expressed in terms of a

Pullum et al. (2007) related an apparent viscosity to a characteristic geometric section

A different modelling approach by Wilson and Sellgren (2006),directly linked to the

3.1 SIMULATED TAILINGS DATA

A simulated a tailings product was characterized by a fully sheared yield stress of

3.2 KAOLIN AND CMC DATA

1.5 N B BEP- Line

Wennberg T and Sellgren A

The 17th international conference on the hydraulic transportation of solids in pipes,

PUMPING EVALUATIONS WITH PASTE TAILINGS

THICKENED CLOSE TO THE SURFACE DISPOSAL AREA

Head ratio: HR = H/H0

0.16 Head water

0.12 Efficiency water

0.02 Coal, C =31%

ANSI/HI 12.1-12.6.standard for rotodynamic (centrifugal) slurry pumps

ENGMAN, M., A.SELLGREN, Å. SUNDQVIST, I.GOLDKUHL AND T. WENNBERG : Users

SCHAAN J, SANDERS, R S GILLIES R G,MCKIBBEN M J,LITZENBERGER C,SUN R-J,SHOOK C

SELLGREN A.AND VÄPPLING L.,Effects of highly concentrated on the performance of centrifgal

Rheological and Depositional characterisation of paste-like tailings slurries

Wennberg T, Sellgren A and Goldkuhl I

Proceedings, Conference in Minerals Engineering, February 3-4 2008, Luleå, Sweden

Thord Wennberg1, Anders Sellgren2, Ingemar Goldkuhl3

2.Objective and scope.

The depositional characterization is mainly a summary of results presented in Engman et al.

Figure 2. Schematic sketch of the disposal area at MultiServ AB in central Sweden.

4.1 Depositional characterization

4.2 Rheological characterisation

('p / 'x) D UgjD

Shear stress (Pa) 100

5.Some case studies

6.2 Thickened tailings deposition

7.Discussion and conclusions

An alternative to a rheologically-based slope modelling approach is now considered and

Robinsky (1999) presented a series of laboratory-scale flume deposition slope results

Feasibility considerations of high density disposal at the LKAB Svappavaara

Wennberg T, Sellgren A and Sundqvist Å

7KHSUHVVXUHJUDGLHQW p / x 3DPIRUWKHIORZLQDKRUL]RQWDOSLSHOLQHLV